- Email: [email protected]
User fees and maternity services in Ethiopia
International Journal of Gynecology and Obstetrics 115 (2011) 310–315
Contents lists available at SciVerse ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
AVERTING MATERNAL DEATH AND DISABILITY
User fees and maternity services in Ethiopia Luwei Pearson a,⁎, Meena Gandhi b, Keseteberhan Admasu c, Emily B. Keyes d a
UNICEF, Ethiopia Save the Children UK, Ethiopia c Federal Ministry of Health, Ethiopia d FHI, Research Triangle Park, USA b
a r t i c l e
i n f o
Keywords: Ethiopia Low-resource country Maternity services Quality of care User fees
a b s t r a c t Objectives: To examine user fees for maternity services and how they relate to provision, quality, and use of maternity services in Ethiopia. Methods: The national assessment of emergency obstetric and newborn care (EmONC) examined user fees for maternity services in 751 health facilities that provided childbirth services in 2008. Results: Overall, only about 6.6% of women gave birth in health facilities. Among facilities that provided delivery care, 68% charged a fee in cash or kind for normal delivery. Health centers should be providing maternity services free of charge (the healthcare financing proclamation), yet 65% still charge for some aspect of care, including drugs and supplies. The average cost for normal and cesarean delivery was US $7.70 and US $51.80, respectively. Nineteen percent of these facilities required payment in advance for treatment of an obstetric emergency. The health facilities that charged user fees had, on average, more delivery beds, deliveries (normal and cesarean), direct obstetric complications treated, and a higher ratio of skilled birth attendants per 1000 deliveries than those that did not charge. The case fatality rate was 3.8% and 7.1% in hospitals that did and did not charge user fees, respectively. Conclusion: Utilization of maternal health services is extremely low in Ethiopia and, although there is a government decree against charging for maternity service, 65% of health centers do charge for some aspects of maternal care. As health facilities are not reimbursed by the government for the costs of maternity services, this loss of revenue may account for the more and better services offered in facilities that continue to charge user fees. User fees are not the only factor that determines utilization in settings where the coverage of maternity services is extremely low. Additional factors include other out-of-pocket payments such as cost of transport and food and lodging for accompanying relatives. It is important to keep quality of care in mind when user fees are under discussion. © 2011 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction User fees are an unaffordable burden for poor households and represent one facet of the social exclusion experienced by the poor. Out-of-pocket payments, which include user fees at public sector facilities, are regressive methods of financing health care, taking a higher proportion of income among poor households than wealthier ones [1]. One challenge to improving maternal health services is that of balancing the need to increase coverage and utilization, particularly among the poor, with the need to improve the quality and financial viability of the health system. Factors related to uptake of health services are many and complex, and vary in countries with low and high coverage of skilled birth attendants (SBAs). A review of the literature on the role of user fees in healthcare services in 5 African countries showed that removing
⁎ Corresponding author. E-mail address: [email protected] (L. Pearson).
them generally has positive effects on utilization of services, but also highlighted issues of quality, workload, provider satisfaction, and implementation [2]. An evaluation of the national free delivery and cesarean policy in Senegal found that there were small increases in utilization for normal deliveries (from 40% to 44% of expected deliveries in the intervention areas from 2004–05) and in (population-based) cesarean rates, which increased from 4.2% to 5.6% [3]. A recent review analyzed 8 case studies using different methods to increase access to obstetric services, including the abolition of user fees, targeted waivers, conditional cash transfers, and insurance schemes. Although service utilization increased with most approaches, concerns remained about the quality of care and rich/poor and urban/ rural equity. There were also concerns about the financial sustainability of these strategies [4]. In Ethiopia, it is clear that the population in the lowest wealth quintile has significantly poorer access to basic health care [5]. A study on the perceptions of user fees for health services showed that fees presented a considerable psychological burden to a family, especially when dealing with unexpected major illnesses. Families usually did not save and were often forced to sell assets in these
0020-7292/$ – see front matter © 2011 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijgo.2011.09.007
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
situations [6]. A more recent study on vulnerable children showed that the cost of illness was a significant factor in tipping families from poverty into extreme vulnerability and exposure to several risks [7]. The fees charged are often not standard and the indirect costs of health care such as transport costs and lodging for the family and food were as much of a burden as the fees themselves. Similarly, the recent National Health Account (NHA) survey round IV found that 4 out of 10 people who had been sick within the 4 weeks preceding the survey did not seek care and by far the most common reason given was affordability. This survey also estimated the national outof-pocket per capita health expenditure to be US $4.15 for outpatients, US $0.46 for inpatients, and US $0.94 for nonhealth expenditures (transport, accommodation, and food etc.). The indirect costs were higher for rural residents than urban. Half of clients walk almost 10 km to get to facilities [8]. The Government of Ethiopia introduced healthcare financing reforms in the most populous regions of the country beginning in 2005. The principles and implementation of the reforms vary from region to region but include user fee retention at facility level, user fee revision, social and community based health insurance schemes, a waiver system for the poorest (identified by fixed criteria), and a standardized list of exempted services that are to be delivered free of charge to everyone at point-of-use at health center and health post level. While expenses to healthcare providers on waiver-related services are reimbursed, those on exempted services are not. These last two components, however, are bold progressive moves toward achieving real equity in healthcare access [9]. The implementation manual for healthcare financing reforms of the Federal Ministry 2005 [9] states that prenatal, delivery, postnatal, and family planning services provided by primary healthcare units (health centers and health posts) should be exempt from payment for all people (regardless of ability to pay), along with tuberculosis treatment, immunization, voluntary counseling and testing (VCT) for HIV, prevention of mother-to-child transmission of HIV (PMTCT), leprosy, and epidemic related services. However, as stated above, the cost of these services is not reimbursed to health centers and must be covered by other revenues. Exemptions do not apply at hospital level. Perceptions of quality of care also play a role in seeking care for maternity services. Studies in both Ethiopia and Tanzania show that a high proportion of people bypass their nearest primary care facilities to seek care in higher-level government facilities or private facilities. Perceived poor quality of care at nearby primary care facilities (as well as the patients’ age, number of children, and use of maternity waiting homes) was significantly associated with bypassing in the Tanzania study [10]. A qualitative study in Ethiopia identified other key factors that affected utilization of maternity services, including a lack of education, low income, lack of awareness of services, distance from services, and health facility related factors [11]. Ethiopia has taken many steps to resolve some of the barriers facing women when seeking maternity services. These include the rapid expansion of health services, moves to quadruple the number of midwives trained and the inclusion of prenatal, delivery, and postnatal care on the list of free services. Indirect cost factors, transport, cultural barriers, and the perceived quality of care have not been sufficiently addressed. 1.1. Purpose We examine the cost and types of user fees charged for maternity services in both government and private hospitals and health centers that offer maternity services in all regions of Ethiopia using data from the emergency obstetric and newborn care (EmONC) assessment in 2008. In addition, we compare the quality and utilization of maternity services between facilities that charged and those that did not.
311
2. Methods The 2008 Ethiopian National EmONC Assessment has been described in detail elsewhere [12]. Briefly, it covered 797 health facilities (750 government, 27 private for profit, 12 NGO, and 8 mission facilities) (Table 1). A total of 751 health facilities provided maternity services (112 hospitals and 639 health centers). Data were collected between October 8, 2008 and January 15, 2009. A total of 84 data collectors, all health professionals, were recruited and trained, then worked in 4-person teams. Hand held GPS units were used to take the location of the health facilities surveyed. Data were reviewed for quality before entry and were double entered in CSPro software (US Census Bureau, Washington DC, USA). The assessment had 10 modules for various aspects of EmONC services, and this paper focuses on the module on user fees related to maternity and EmONC. In all health facilities surveyed, we asked staff about types and amounts of user fees for maternity services including card fees (registration fee required before consultation), consultation fees, charges for delivery services, lab services, and essential commodities for maternity service, such as drugs and supplies. We also asked if payment is required before a woman can receive treatment, including treatment for obstetric emergencies. We analyzed the data to identify differences in quantity and quality of maternity services between the facilities that charge user fees and those that do not.
3. Results 3.1. Costs of EmONC Services 3.1.1. Card fees and payment policies In Ethiopia, 54% of health facilities report that they require a card fee prior to providing services. The average charge for the card was US $0.40 in government hospitals and US $2.80 in the nongovernment hospitals. Fees were lower in health centers/clinics. Fee schedules were posted and visible in 29% of facilities (data not shown). Among facilities (private and government) providing delivery services, 68% (85% in hospitals and 66% in health centers) charged a fee for normal delivery or required women to buy supplies for a normal delivery. Percentages were slightly higher in nongovernment facilities than government facilities. One-fifth of facilities with delivery services required payment in advance for an obstetric emergency. Three-quarters of nongovernment and 30% of government hospitals required payment in advance. About 1 in 5 health centers/clinics required payment before treatment for an emergency (Table 1).
Table 1 Percentage of facilities that require women to purchase supplies and that require payment before emergency among facilities that perform deliveries, Ethiopia, 2008. Charge fee or require woman to buy supplies for normal delivery, % National Facility type Hospital Government Other Health center/clinic Government Other
Require payment prior to treatment for obs/gyn emergency, % a
Total number of facilities that perform deliveries
68
19
751
85 83 91 66 65 71
38 30 73 16 16 21
112 90 22 639 625 14
a
a Two health centers did not answer and were excluded from denominators in these columns.
312
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
Table 2 Availability and average charges for various obstetric and gynecologic services by facility type/sector, Ethiopia, 2008. Normal delivery
National Hospitals Government Other Health centers Government Other
No. of facilities that offer service
% that charge
751 112 90 22 639 625 14
26 77 73 95 17 16 64
Cesarean delivery a
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
7.70 14.40 4.10 46.40 2.30 1.90 7.40
93 87 66 21 6 5 1
80 85 82 95 17 0 100
Assisted vaginal delivery (AVD) b
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
51.80 51.10 13.50 158.40 [102.00]5 [102.00] e
273 99 80 19 174 168 6
34 76 74 83 10 9 50
c
Safe abortion
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
15.00 17.40 5.60 62.30 4.60 1.60 18.50
356 95 80 15 261 254 7
47 87 85 100 32 32 57
d
Average cost among those that charge, US$ 6.60 10.00 4.80 34.90 3.50 3.20 9.30
Note 1: Averages based on fewer than 10 observations appear in brackets [ ]. Note 2: Facilities that answered “did not know” to cost questions were included in the number of facilities that offer the service. However, they were excluded from the calculation of the percent that charge. Facilities that had conflicting answers about whether they provided the service at all (across modules), but where the key variable used to determine provision of service says that they did provide service, were included as a facility that provided service and were included in the calculation of the percent charging (assumed to not charge). This provides a more conservative estimate. a Two facilities with conflicting answers about provision of service (both public). b One facility with conflicting answers about provision of service (public). c Twelve facilities responded “do not know the cost” (8 public health centers, 3 public hospitals, 1 private hospital). d Seven facilities responded “do not know the cost” (5 public health centers, 1 public hospital, 1 public hospital). e This average was based on one response from a private health center.
3.1.2. Costs for selected services, supplies, and essential drugs Table 2 shows that of facilities offering the specific service, 26% charged for normal delivery, 80% for a cesarean, 34% for assisted vaginal delivery, and 47% for safe abortion. Across all services, as expected, hospitals were more likely to charge than health centers, and private for-profit and nonprofit facilities more likely to charge than government facilities. Among facilities that charged for services the most costly service was cesarean delivery. Government hospitals charged on average US $4.10 for a normal delivery, US $4.80 for a safe abortion, US $5.60 for assisted vaginal delivery, and US $13.50 for a cesarean delivery. The nongovernment facilities were more likely to charge for any of these services and to charge more than the government sector (Table 2). Nearly two-thirds of the 751 facilities charged for extra supplies and a large proportion of these were government health centers. The charges were, on average, US $0.80 for gloves, syringes, and needles. Seventy-five percent charged an average of US $1.80 for intravenous fluids and catheters. Important lifesaving obstetric drugs, including oxytocin, penicillin, and gentamicin required an average payment of US $0.80, US $2.00, and US $1.10, respectively. Facilities reported they were more likely to charge for gloves, intravenous fluids, or medications than for beds or food (data not shown). Again, there is a tendency for government facilities to be less likely to charge and to charge less than other facilities (Table 3). 3.1.3. Fee waiver Comprehensive obstetric care services are only available at hospital level, where exemptions do not apply. In addition, exemptions are
not uniformly implemented at health center level. For these reasons, waivers are particularly important. Nationally, 68% of all health facilities reported a formal system to waive fees for poor women, 13% had an informal system, and approximately 19% reported no system. Both hospitals (65%) and health centers (69%) reported that they implement a formal waiver system. When examined by management sector, we see that most government hospitals (76%) had a formal system, but only 23% of nongovernment hospitals had a formal system. The pattern is repeated for health centers/clinics [12]. 3.2. Utilization of EmONC services 3.2.1. Overall utilization The expected annual number of births in Ethiopia is 2.54 million (total population of 74 million and a crude birth rate of 35.7 in 2008). In the 751 facilities surveyed, the total number of deliveries was 174 561, which is 6.6% of the total expected births, and 43% of these took place in health centers. The data shown below (and Table 4), therefore, refer only to the 6.6% of the women who delivered in health facilities. 3.2.2. Volume of service The hospitals and health centers that charged user fees had a higher volume of services provided in general. During Ethiopia's fiscal year 2001 (July 2007 to June 2008 as the western calendar is not used), for facilities that did and did not charge, the mean annual number of deliveries attended was 951 and 528 for hospitals and 130 and 96 for health centers, respectively. The average number of cesarean
Table 3 Percentage of facilities that charged for essential drugs and commodities and average charge by facility type/sector (among facilities that perform deliveries), Ethiopia, 2008. Intavenous fluids, catheter
Total number of facilities
Gloves, syringe, needles
Oxytocin
Penicillin
Gentamicin
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
National
751
61
0.80
75
1.80
43
0.80
69
2.00
77
1.10
Facility type Hospital Government Other Health Center/clinic Government Other
112 90 22 635 621 14
75 74 75 58 58 50
0.90 0.90 0.60 0.70 0.70 1.10
81 81 81 74 74 69
2.90 2.70 3.70 1.70 1.70 1.80
74 73 79 38 38 38
1.90 2.10 1.00 0.50 0.50 0.90
75 78 63 68 68 58
3.50 3.40 4.10 1.70 1.70 2.40
83 82 84 76 76 55
1.50 1.20 2.60 1.10 1.10 1.10
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
313
Table 4 Characteristics of facilities by presence of user fees for normal delivery and facility type, Ethiopia, 2008. Hospitals
Managing organization Government Private, for-profit Private, not for profit (NGO or religious) Volume of services Deliveries (average annual) Cesarean deliveries (average annual) Number of all direct obstetric complications treated (average annual) Condition of basic infrastructure Number of delivery beds and couches (average) Water connection (piped water) Electricity connection (grid and generator) Ambulance available (any motor vehicle) Communication facilities (any on-site) Human resources SBA available 24 hours per day, 7 days per week Ratio of SBAs per 1000 institutional deliveries Total number of SBAs (average) Total number of all health workers (average) Quality of care Oxytocin available Full package of comprehensive EmONC Availability of resuscitators for newborn asphyxia Use of partograph in last 3 months Case fatality rate of direct obstetric complications
Health centers
Charge (n=95)
Do not charge (n=17)
Charge (n=418)
Do not charge (n=219)
80% 16% 5%
88% 6% 6%
98% 0% 2%
98% 0% 2%
951 158 258
528 99 219
130 28
96 17
26.5 96% 85% 85% 94%
17.3 94% 59% 88% 94%
4.4 70% 17% 42% 58%
4.1 70% 29% 43% 61%
96% 59.5 12.5 72.9
94% 46.5 12.4 76.8
56% 50.3 2.2 11.6
53% 40.1 1.8 11
95% 58% 77% 26% 3.8%
82% 53% 71% 29% 7.1%
55% 35% 27% 1.2%
58% 34% 24% 0.4%
Abbreviations: SBA, skilled birth attendant defined as an obstetrician/gynecologist, medical doctor, health officer, or midwife; EmONC, emergency obstetric and newborn care.
deliveries conducted was 158 and 99 for hospitals that did and did not charge user fees. The average number of direct obstetric complications treated was 258 and 219 for hospitals, and 28 and 17 for health centers that did and did not charge user fees. (Table 4). 3.3. Condition of basic infrastructure Hospitals that charged user fees had more maternity beds than those that did not charge (26.5 vs 17.3) and were much more likely to have the most reliable source of electricity (i.e. electric grid with backup generator). However, in terms of other basic infrastructure items, there was little difference between health facilities that did and did not charge user fees. Nearly all the hospitals reported having piped water, while only 70% of the health centers did. Over 80% of hospitals and 40% of health centers had a functioning motor vehicle for emergency referral. Nearly all the government hospitals and 58% of health centers had on-site communication facilities. 3.3.1. Human resources There was little difference in the availability of SBAs in health facilities that did and did not charge user fees. Almost all hospitals, and 53% and 56% of health centers that did and did not charge user fees had SBAs. However, the average ratio of SBAs to 1000 institutional deliveries did differ significantly: 60 per 1000 compared with 47 per 1000 for hospitals, and 50 per 1000 compared with 40 per 1000 for health centers that did and did not charge user fees, respectively. 3.3.2. Quality of service There was little difference in terms of quality of care provided between the health facilities that did and did not charge user fees. Just over half (58%) of hospitals that charged user fees provided the full package of EmONC compared with 53% of those that did not. Similar findings were seen for all the standard parameters of quality used in the emergency obstetric care assessment (Table 4): use of a partograph to monitor labor, availability of oxytocin, managing direct obstetric complications, and the availability of newborn resuscitation.
The case fatality rate (CFR) for direct obstetric complications varied significantly at 3.8% for hospitals that charged user fees compared with 7.1% for those that did not, and 1.2% for the health centers that charged user fees and 0.4% for those that did not. These contradictory findings require caution as they do not consider the condition of the patients going to the free facilities, the extent of underreporting of complications or deaths, and high rates of referral out, for example. The findings may also be due to sampling errors since there were very few severe obstetric complications managed at health centers. 4. Discussion This assessment of maternity services in Ethiopia has some limitations, which are described in detail elsewhere [12]. Facility records of deliveries, obstetric complications, cesarean deliveries, and deaths were often incomplete. Deaths, in particular, can be omitted or classified inaccurately, in part because maternal deaths from indirect causes are less likely to be found in the maternity or gynecology wards. Under-recording of complications (and deaths) will affect the direct obstetric case fatality rate. Assessment of equipment, supplies and drugs was encouraged but given the very long lists of these articles, not all items were observed. Although designed as a census of hospitals and health centers, ensuring that all hospitals, health centers, and higher clinics were visited proved challenging. The list of facilities registered in the Ministry database was incomplete. The coverage of facilities was incomplete in some regions. For security reasons, only 11 of the 26 facilities were visited in Somali Region, although most hospitals were visited. In Addis Ababa, the original list of facilities obtained from the local Regional Health Bureau was restricted to those that were licensed at that time. After the conclusion of data collection it was learned that some private clinics and hospitals were not included. Also, in Addis Ababa, some facilities were not forthcoming with data for some of the modules, and these data were not included in the analysis. Thus, it is likely that the indicators that reflect availability and utilization are conservative estimates.
314
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
The national emergency obstetric care assessment clearly shows severe underutilization of maternal health services (6.6% of the total expected annual births) and there are many reasons for this, including geographic distance, poor roads, lack of transport, perceptions of poor quality and availability of care, and economic constraints. A 2005 community- based survey by the safe motherhood program of Ethiopia found that decision-making about where childbirth takes place involves many members of the family and community, who retain control of the process and the outcome of births. Timely care seeking depends on the knowledge, understanding, and financial means of the household—usually controlled by the husband. Distance, cost, and considerable support for the cultural practices around birth hinder the use of modern health services [13]. Still, overall poverty seems to be the biggest reason for people not to seek care when sick. The average income in Ethiopia was US $0.68 per day in 2010 leaving little flexibility to manage unanticipated expenses. Other factors in addition to costs are physical access including referral and perceived quality of care. These should also be addressed as the removal of user fees is implemented [14]. Since there is already a policy in Ethiopia that maternity services (among others) at health centers should be free of charge, consideration needs to be given to regulation and enforcement. The waiver system is used to provide free services to the poorest at all levels, including hospital. Poorer households are preidentified to receive all health services free and are given certificates valid for 3 years. A recent review in Amhara region found that the revised fee waiver system had improved timely access for the poor to health services, but reimbursement to the facilities for waiver patients by the district offices varied from 25% to 100%. Some health centers received only US $0.40 per patient, which is much lower than the costs of a normal delivery. Problems were also identified with the fee waiver system such as targeting errors, limits to numbers of beneficiaries, and incomplete exemptions [15]. The recent National Health Accounts (NHA) IV Household and Utilization survey discovered that the poorest quintile of the population actually has the lowest percentage of fee waiver beneficiaries (data not shown) [8]. This provides evidence that the poorest are likely being missed in waiver targeting or facing other barriers to access of available services. The waiver system currently in operation in Ethiopia seems to be operating in the majority of government facilities, but there remain significant issues in terms of targeting the waivers to the neediest. More efficient systems of community-based targeting could be piloted during the roll out of the social and community-based health insurance systems in the country. The NHA also revealed that the national health expenditure (NHE) increased from US $522 million in 2004/05 to over US $1.2 billion in 2007/08. Per capita NHE also grew substantially, more than doubling from US $7.14 per capita per annum in 2004/05 to US $16.09 in 2007/08. This figure, however, includes funding received from all sources, including patients’ out- of-pocket expenditures. The major funding sources for reproductive health care in particular were international donors (36%), regional and local governments (28%), and households (25%) [8]. According to the Demographic Health Survey 2011 [5], the national average of institutional births is 10%: 51% for urban and 4% for rural areas. It is likely that the majority of users of maternity services have greater capacity to pay and are more motivated to seek care and therefore less influenced by the price. This is reflected in the findings that facilities that charge user fees see more patients on average than those that do not. The exemption policy regarding maternal services in place at primary healthcare unit level removes one of the key barriers to appropriate care seeking; if it is implemented correctly. A recent study in Ghana showed a nonstatistically significant decline in deliveryrelated deaths following the implementation of the exemption policy. Nor was there evidence that the severity of delivery-related obstetric conditions was significantly less before and after the exemption. In
contrast, interventions to improve access to care at the Juaben Teaching Health Center in the Ashanti Region of Ghana for example, led to a 3fold increase among women with complications seeking care and a 67% drop in referrals for treatment [16]. Our study showed that free maternity care is implemented in only a minority of facilities, with 65% of government health centers charging for services, supplies, drugs, or everything. Unlike waivers, there is no formal system of reimbursement of exempted services and health centers are expected to absorb these costs. At the same time, health centers are encouraged to raise and retain their own revenue through user fees for nonexempt services allowing them to improve their facilities. Some services on the exempted list are actually provided free of charge at the point of use (malaria treatment and immunizations, for example) and are those that are funded by international donors. Although the quality of many service parameters varies between facilities that charge and those that do not, it is clear that hospitals that charge are better equipped. One critical difference is the more favorable ratio of SBAs in facilities that charge compared with those that do not. This affects quality of care (for example as reflected by case fatality rates) and patients’ perceptions of quality of care (as reflected by utilization rates). The mandate to provide free maternity services in health centers to all women and the waiver system for hospital care for the poorest deprives health facilities of a significant source of revenue. In addition, the level of reimbursement of waiver services is not standardized and is insufficient to cover the costs of care. Providing emergency obstetric care at no cost to poor women is an important step forward in reducing the number of women who die in childbirth. However, the economics is complex. If care providers are not compensated in some way for this loss of revenue the quality of care provided will inevitably suffer. This is no small challenge. Health economists have developed such methods as low cost insurance schemes, increasing fees charged to patients better able to pay, installment payments, and voucher schemes. Governments must consider alternative sources of revenue to providers when considering mandates to provide free care to a part of the population. 5. Recommendations Generally speaking, when user fees cannot be removed completely, a targeted fee waiver for the poor and an exemption system for women suffering from obstetric complications, particularly those unable to pay, should be considered. Increasing coverage of health insurance can also increase availability of affordable maternity services. These are important steps to increase use of lifesaving services and to achieve MDGs 4 and 5. When progressive policies such as the fee waiver and exemption system are in place, consistent implementation is needed to ensure that the poorest are adequately targeted. Either all exempted services should be reimbursable to health providers or other means of cost recovery should be developed. Balancing these measures with maintaining high quality of care should not be overlooked. Addressing other barriers to access such as opportunity costs and distance remain a challenge. To enable free maternity services, health facilities should be adequately reimbursed for the essential supplies, commodities, and services rendered. The user fees can be compensated in different ways such as providing clean delivery kits, vouchers, or direct cash support based on the services provided. Communities should be informed of the eligibility criteria for fee waivers and services exempted from user fees. This would create demand and improve uptake of available services. For Ethiopia specifically, there is room to increase the government's allocation to the health sector. The Growth and Transformation Plan (2010–2015) of the Government of Ethiopia plans to double the government allocation to the health sector, which is a great opportunity to increase investment in the health system and maternal and newborn health in particular. Since waivers are issued
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
and exempt services are reimbursed by Woredas (districts), there should be sufficient budget allocation from the central and regional governments for these activities. In the short term, financial support from development partners is important to bridge the financial gap. Acknowledgments The authors are grateful to the Ethiopian Federal Ministry of Health and Regional Health Bureau teams for their support. We recognize the effort of the many individuals at UNICEF, UNFPA, WHO, BETA, and AMDD who contributed to this Assessment. The assessment was jointly funded by UNICEF, UNFPA, and WHO. AMDD provided technical assistance. Conflict of interest None. References [1] Van Doorslaer E, Wagstaff A. Equity in the finance of health care: Methods and findings. In: Van Doorslaer E, Wagstaff A, Rutten F, editors. Equity in the finance and delivery of health care: An international perspective. New York: Oxford University Press; 1993. [2] Ridde V, Morestin F. A scoping review of the literature on the abolition of user fees in health care services in Africa. Health Policy Plan 2011;26(1):1–11. [3] Witter S, Mbengue D, Moreira I, De Brouwere V. The national free delivery and caesarean policy in Senegal: Evaluating process and outcomes. Health Policy Plan 2010;25(5):384–92.
315
[4] Richard F, Witter S, de Brouwere V. Innovative approaches to reducing financial barriers to obstetric care in low-income countries. Am J Public Health 2010;100(10): 1845–52. [5] Central Statistical Agency. Ethiopia Demographic and Health Survey 2011 (Preliminary report). Addis Ababa, Ethiopia: Central Statistical Agency; 2011. [6] Barnett I, Tefera B. Poor households’ experiences and perceptions of user fees for healthcare: a mixed method study from Ethiopia. Young Lives International Study of Childhood Poverty. Working paper no. 59. June 2010. Available at: http:// www.younglives.org.uk/our-publications/working-papers/poor-households2019experiences-and-perception-of-user-fees-for-healthcare. [7] 2nd Oak Foundation and Young Lives study consultation on vulnerable children, 6th December 2010, Addis Ababa. Unpublished. [8] Federal Ministry of Health. National Health Accounts IV: Household Health Service Utilisation and Expenditure Survey report; 2010. Addis Ababa, Ethiopia. [9] Federal Ministry of Health. Implementation manual for health care financing reports; 2010. Addis Ababa, Ethiopia. [10] Kruk ME. Bypassing primary care facilities for childbirth: a population-based study in rural Tanzania. Health Policy Plan 2009;24(4):279–88. [11] Kumbi S, Debebe A, Assefa G. Reasons for underutilization of emergency obstetric series in three Woredas of West Shoa Zone, Oromia region, Ethiopia: a qualitative study. Paper presented at the 12th Annual Conference of the Ethiopian Society of Obstetricians and Gynecologists; 2002. [12] Federal Ministry of Health. National Baseline Assessment for Emergency Obstetric and Newborn Care; 2008. Addis Ababa, Ethiopia. [13] Worren C. Care seeking for maternal health: challenges remain for poor women. Ethiop J Health Dev 2010;24:87–129. [14] Kruk ME, Mbaruku G, Rockers PC, Galea S. User fee exemptions are not enough: out-of-pocket payments for 'free' delivery services in rural Tanzania. Trop Med Int Health 2008;13(12):1442–51. [15] Save the Children UK. Situation analysis of health care financing reform. Implementation in five districts of south Wollo zone of Amhara region; 2009. Ethiopia: Successes and challenges. London: Save the Children. [16] Bosu W, Bell JS, Armar-Klemesu M, Tornui JA. Effect of delivery care user fee exemption policy on institutional maternal deaths in the central and volta regions of Ghana. Ghana Med J 2007;41(3):118–24.
Contents lists available at SciVerse ScienceDirect
International Journal of Gynecology and Obstetrics journal homepage: www.elsevier.com/locate/ijgo
AVERTING MATERNAL DEATH AND DISABILITY
User fees and maternity services in Ethiopia Luwei Pearson a,⁎, Meena Gandhi b, Keseteberhan Admasu c, Emily B. Keyes d a
UNICEF, Ethiopia Save the Children UK, Ethiopia c Federal Ministry of Health, Ethiopia d FHI, Research Triangle Park, USA b
a r t i c l e
i n f o
Keywords: Ethiopia Low-resource country Maternity services Quality of care User fees
a b s t r a c t Objectives: To examine user fees for maternity services and how they relate to provision, quality, and use of maternity services in Ethiopia. Methods: The national assessment of emergency obstetric and newborn care (EmONC) examined user fees for maternity services in 751 health facilities that provided childbirth services in 2008. Results: Overall, only about 6.6% of women gave birth in health facilities. Among facilities that provided delivery care, 68% charged a fee in cash or kind for normal delivery. Health centers should be providing maternity services free of charge (the healthcare financing proclamation), yet 65% still charge for some aspect of care, including drugs and supplies. The average cost for normal and cesarean delivery was US $7.70 and US $51.80, respectively. Nineteen percent of these facilities required payment in advance for treatment of an obstetric emergency. The health facilities that charged user fees had, on average, more delivery beds, deliveries (normal and cesarean), direct obstetric complications treated, and a higher ratio of skilled birth attendants per 1000 deliveries than those that did not charge. The case fatality rate was 3.8% and 7.1% in hospitals that did and did not charge user fees, respectively. Conclusion: Utilization of maternal health services is extremely low in Ethiopia and, although there is a government decree against charging for maternity service, 65% of health centers do charge for some aspects of maternal care. As health facilities are not reimbursed by the government for the costs of maternity services, this loss of revenue may account for the more and better services offered in facilities that continue to charge user fees. User fees are not the only factor that determines utilization in settings where the coverage of maternity services is extremely low. Additional factors include other out-of-pocket payments such as cost of transport and food and lodging for accompanying relatives. It is important to keep quality of care in mind when user fees are under discussion. © 2011 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction User fees are an unaffordable burden for poor households and represent one facet of the social exclusion experienced by the poor. Out-of-pocket payments, which include user fees at public sector facilities, are regressive methods of financing health care, taking a higher proportion of income among poor households than wealthier ones [1]. One challenge to improving maternal health services is that of balancing the need to increase coverage and utilization, particularly among the poor, with the need to improve the quality and financial viability of the health system. Factors related to uptake of health services are many and complex, and vary in countries with low and high coverage of skilled birth attendants (SBAs). A review of the literature on the role of user fees in healthcare services in 5 African countries showed that removing
⁎ Corresponding author. E-mail address: [email protected] (L. Pearson).
them generally has positive effects on utilization of services, but also highlighted issues of quality, workload, provider satisfaction, and implementation [2]. An evaluation of the national free delivery and cesarean policy in Senegal found that there were small increases in utilization for normal deliveries (from 40% to 44% of expected deliveries in the intervention areas from 2004–05) and in (population-based) cesarean rates, which increased from 4.2% to 5.6% [3]. A recent review analyzed 8 case studies using different methods to increase access to obstetric services, including the abolition of user fees, targeted waivers, conditional cash transfers, and insurance schemes. Although service utilization increased with most approaches, concerns remained about the quality of care and rich/poor and urban/ rural equity. There were also concerns about the financial sustainability of these strategies [4]. In Ethiopia, it is clear that the population in the lowest wealth quintile has significantly poorer access to basic health care [5]. A study on the perceptions of user fees for health services showed that fees presented a considerable psychological burden to a family, especially when dealing with unexpected major illnesses. Families usually did not save and were often forced to sell assets in these
0020-7292/$ – see front matter © 2011 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijgo.2011.09.007
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
situations [6]. A more recent study on vulnerable children showed that the cost of illness was a significant factor in tipping families from poverty into extreme vulnerability and exposure to several risks [7]. The fees charged are often not standard and the indirect costs of health care such as transport costs and lodging for the family and food were as much of a burden as the fees themselves. Similarly, the recent National Health Account (NHA) survey round IV found that 4 out of 10 people who had been sick within the 4 weeks preceding the survey did not seek care and by far the most common reason given was affordability. This survey also estimated the national outof-pocket per capita health expenditure to be US $4.15 for outpatients, US $0.46 for inpatients, and US $0.94 for nonhealth expenditures (transport, accommodation, and food etc.). The indirect costs were higher for rural residents than urban. Half of clients walk almost 10 km to get to facilities [8]. The Government of Ethiopia introduced healthcare financing reforms in the most populous regions of the country beginning in 2005. The principles and implementation of the reforms vary from region to region but include user fee retention at facility level, user fee revision, social and community based health insurance schemes, a waiver system for the poorest (identified by fixed criteria), and a standardized list of exempted services that are to be delivered free of charge to everyone at point-of-use at health center and health post level. While expenses to healthcare providers on waiver-related services are reimbursed, those on exempted services are not. These last two components, however, are bold progressive moves toward achieving real equity in healthcare access [9]. The implementation manual for healthcare financing reforms of the Federal Ministry 2005 [9] states that prenatal, delivery, postnatal, and family planning services provided by primary healthcare units (health centers and health posts) should be exempt from payment for all people (regardless of ability to pay), along with tuberculosis treatment, immunization, voluntary counseling and testing (VCT) for HIV, prevention of mother-to-child transmission of HIV (PMTCT), leprosy, and epidemic related services. However, as stated above, the cost of these services is not reimbursed to health centers and must be covered by other revenues. Exemptions do not apply at hospital level. Perceptions of quality of care also play a role in seeking care for maternity services. Studies in both Ethiopia and Tanzania show that a high proportion of people bypass their nearest primary care facilities to seek care in higher-level government facilities or private facilities. Perceived poor quality of care at nearby primary care facilities (as well as the patients’ age, number of children, and use of maternity waiting homes) was significantly associated with bypassing in the Tanzania study [10]. A qualitative study in Ethiopia identified other key factors that affected utilization of maternity services, including a lack of education, low income, lack of awareness of services, distance from services, and health facility related factors [11]. Ethiopia has taken many steps to resolve some of the barriers facing women when seeking maternity services. These include the rapid expansion of health services, moves to quadruple the number of midwives trained and the inclusion of prenatal, delivery, and postnatal care on the list of free services. Indirect cost factors, transport, cultural barriers, and the perceived quality of care have not been sufficiently addressed. 1.1. Purpose We examine the cost and types of user fees charged for maternity services in both government and private hospitals and health centers that offer maternity services in all regions of Ethiopia using data from the emergency obstetric and newborn care (EmONC) assessment in 2008. In addition, we compare the quality and utilization of maternity services between facilities that charged and those that did not.
311
2. Methods The 2008 Ethiopian National EmONC Assessment has been described in detail elsewhere [12]. Briefly, it covered 797 health facilities (750 government, 27 private for profit, 12 NGO, and 8 mission facilities) (Table 1). A total of 751 health facilities provided maternity services (112 hospitals and 639 health centers). Data were collected between October 8, 2008 and January 15, 2009. A total of 84 data collectors, all health professionals, were recruited and trained, then worked in 4-person teams. Hand held GPS units were used to take the location of the health facilities surveyed. Data were reviewed for quality before entry and were double entered in CSPro software (US Census Bureau, Washington DC, USA). The assessment had 10 modules for various aspects of EmONC services, and this paper focuses on the module on user fees related to maternity and EmONC. In all health facilities surveyed, we asked staff about types and amounts of user fees for maternity services including card fees (registration fee required before consultation), consultation fees, charges for delivery services, lab services, and essential commodities for maternity service, such as drugs and supplies. We also asked if payment is required before a woman can receive treatment, including treatment for obstetric emergencies. We analyzed the data to identify differences in quantity and quality of maternity services between the facilities that charge user fees and those that do not.
3. Results 3.1. Costs of EmONC Services 3.1.1. Card fees and payment policies In Ethiopia, 54% of health facilities report that they require a card fee prior to providing services. The average charge for the card was US $0.40 in government hospitals and US $2.80 in the nongovernment hospitals. Fees were lower in health centers/clinics. Fee schedules were posted and visible in 29% of facilities (data not shown). Among facilities (private and government) providing delivery services, 68% (85% in hospitals and 66% in health centers) charged a fee for normal delivery or required women to buy supplies for a normal delivery. Percentages were slightly higher in nongovernment facilities than government facilities. One-fifth of facilities with delivery services required payment in advance for an obstetric emergency. Three-quarters of nongovernment and 30% of government hospitals required payment in advance. About 1 in 5 health centers/clinics required payment before treatment for an emergency (Table 1).
Table 1 Percentage of facilities that require women to purchase supplies and that require payment before emergency among facilities that perform deliveries, Ethiopia, 2008. Charge fee or require woman to buy supplies for normal delivery, % National Facility type Hospital Government Other Health center/clinic Government Other
Require payment prior to treatment for obs/gyn emergency, % a
Total number of facilities that perform deliveries
68
19
751
85 83 91 66 65 71
38 30 73 16 16 21
112 90 22 639 625 14
a
a Two health centers did not answer and were excluded from denominators in these columns.
312
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
Table 2 Availability and average charges for various obstetric and gynecologic services by facility type/sector, Ethiopia, 2008. Normal delivery
National Hospitals Government Other Health centers Government Other
No. of facilities that offer service
% that charge
751 112 90 22 639 625 14
26 77 73 95 17 16 64
Cesarean delivery a
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
7.70 14.40 4.10 46.40 2.30 1.90 7.40
93 87 66 21 6 5 1
80 85 82 95 17 0 100
Assisted vaginal delivery (AVD) b
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
51.80 51.10 13.50 158.40 [102.00]5 [102.00] e
273 99 80 19 174 168 6
34 76 74 83 10 9 50
c
Safe abortion
Average cost among those that charge, US$
No. of facilities that offer service
% that charge
15.00 17.40 5.60 62.30 4.60 1.60 18.50
356 95 80 15 261 254 7
47 87 85 100 32 32 57
d
Average cost among those that charge, US$ 6.60 10.00 4.80 34.90 3.50 3.20 9.30
Note 1: Averages based on fewer than 10 observations appear in brackets [ ]. Note 2: Facilities that answered “did not know” to cost questions were included in the number of facilities that offer the service. However, they were excluded from the calculation of the percent that charge. Facilities that had conflicting answers about whether they provided the service at all (across modules), but where the key variable used to determine provision of service says that they did provide service, were included as a facility that provided service and were included in the calculation of the percent charging (assumed to not charge). This provides a more conservative estimate. a Two facilities with conflicting answers about provision of service (both public). b One facility with conflicting answers about provision of service (public). c Twelve facilities responded “do not know the cost” (8 public health centers, 3 public hospitals, 1 private hospital). d Seven facilities responded “do not know the cost” (5 public health centers, 1 public hospital, 1 public hospital). e This average was based on one response from a private health center.
3.1.2. Costs for selected services, supplies, and essential drugs Table 2 shows that of facilities offering the specific service, 26% charged for normal delivery, 80% for a cesarean, 34% for assisted vaginal delivery, and 47% for safe abortion. Across all services, as expected, hospitals were more likely to charge than health centers, and private for-profit and nonprofit facilities more likely to charge than government facilities. Among facilities that charged for services the most costly service was cesarean delivery. Government hospitals charged on average US $4.10 for a normal delivery, US $4.80 for a safe abortion, US $5.60 for assisted vaginal delivery, and US $13.50 for a cesarean delivery. The nongovernment facilities were more likely to charge for any of these services and to charge more than the government sector (Table 2). Nearly two-thirds of the 751 facilities charged for extra supplies and a large proportion of these were government health centers. The charges were, on average, US $0.80 for gloves, syringes, and needles. Seventy-five percent charged an average of US $1.80 for intravenous fluids and catheters. Important lifesaving obstetric drugs, including oxytocin, penicillin, and gentamicin required an average payment of US $0.80, US $2.00, and US $1.10, respectively. Facilities reported they were more likely to charge for gloves, intravenous fluids, or medications than for beds or food (data not shown). Again, there is a tendency for government facilities to be less likely to charge and to charge less than other facilities (Table 3). 3.1.3. Fee waiver Comprehensive obstetric care services are only available at hospital level, where exemptions do not apply. In addition, exemptions are
not uniformly implemented at health center level. For these reasons, waivers are particularly important. Nationally, 68% of all health facilities reported a formal system to waive fees for poor women, 13% had an informal system, and approximately 19% reported no system. Both hospitals (65%) and health centers (69%) reported that they implement a formal waiver system. When examined by management sector, we see that most government hospitals (76%) had a formal system, but only 23% of nongovernment hospitals had a formal system. The pattern is repeated for health centers/clinics [12]. 3.2. Utilization of EmONC services 3.2.1. Overall utilization The expected annual number of births in Ethiopia is 2.54 million (total population of 74 million and a crude birth rate of 35.7 in 2008). In the 751 facilities surveyed, the total number of deliveries was 174 561, which is 6.6% of the total expected births, and 43% of these took place in health centers. The data shown below (and Table 4), therefore, refer only to the 6.6% of the women who delivered in health facilities. 3.2.2. Volume of service The hospitals and health centers that charged user fees had a higher volume of services provided in general. During Ethiopia's fiscal year 2001 (July 2007 to June 2008 as the western calendar is not used), for facilities that did and did not charge, the mean annual number of deliveries attended was 951 and 528 for hospitals and 130 and 96 for health centers, respectively. The average number of cesarean
Table 3 Percentage of facilities that charged for essential drugs and commodities and average charge by facility type/sector (among facilities that perform deliveries), Ethiopia, 2008. Intavenous fluids, catheter
Total number of facilities
Gloves, syringe, needles
Oxytocin
Penicillin
Gentamicin
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
Percent charging, %
Average charge, US$
National
751
61
0.80
75
1.80
43
0.80
69
2.00
77
1.10
Facility type Hospital Government Other Health Center/clinic Government Other
112 90 22 635 621 14
75 74 75 58 58 50
0.90 0.90 0.60 0.70 0.70 1.10
81 81 81 74 74 69
2.90 2.70 3.70 1.70 1.70 1.80
74 73 79 38 38 38
1.90 2.10 1.00 0.50 0.50 0.90
75 78 63 68 68 58
3.50 3.40 4.10 1.70 1.70 2.40
83 82 84 76 76 55
1.50 1.20 2.60 1.10 1.10 1.10
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
313
Table 4 Characteristics of facilities by presence of user fees for normal delivery and facility type, Ethiopia, 2008. Hospitals
Managing organization Government Private, for-profit Private, not for profit (NGO or religious) Volume of services Deliveries (average annual) Cesarean deliveries (average annual) Number of all direct obstetric complications treated (average annual) Condition of basic infrastructure Number of delivery beds and couches (average) Water connection (piped water) Electricity connection (grid and generator) Ambulance available (any motor vehicle) Communication facilities (any on-site) Human resources SBA available 24 hours per day, 7 days per week Ratio of SBAs per 1000 institutional deliveries Total number of SBAs (average) Total number of all health workers (average) Quality of care Oxytocin available Full package of comprehensive EmONC Availability of resuscitators for newborn asphyxia Use of partograph in last 3 months Case fatality rate of direct obstetric complications
Health centers
Charge (n=95)
Do not charge (n=17)
Charge (n=418)
Do not charge (n=219)
80% 16% 5%
88% 6% 6%
98% 0% 2%
98% 0% 2%
951 158 258
528 99 219
130 28
96 17
26.5 96% 85% 85% 94%
17.3 94% 59% 88% 94%
4.4 70% 17% 42% 58%
4.1 70% 29% 43% 61%
96% 59.5 12.5 72.9
94% 46.5 12.4 76.8
56% 50.3 2.2 11.6
53% 40.1 1.8 11
95% 58% 77% 26% 3.8%
82% 53% 71% 29% 7.1%
55% 35% 27% 1.2%
58% 34% 24% 0.4%
Abbreviations: SBA, skilled birth attendant defined as an obstetrician/gynecologist, medical doctor, health officer, or midwife; EmONC, emergency obstetric and newborn care.
deliveries conducted was 158 and 99 for hospitals that did and did not charge user fees. The average number of direct obstetric complications treated was 258 and 219 for hospitals, and 28 and 17 for health centers that did and did not charge user fees. (Table 4). 3.3. Condition of basic infrastructure Hospitals that charged user fees had more maternity beds than those that did not charge (26.5 vs 17.3) and were much more likely to have the most reliable source of electricity (i.e. electric grid with backup generator). However, in terms of other basic infrastructure items, there was little difference between health facilities that did and did not charge user fees. Nearly all the hospitals reported having piped water, while only 70% of the health centers did. Over 80% of hospitals and 40% of health centers had a functioning motor vehicle for emergency referral. Nearly all the government hospitals and 58% of health centers had on-site communication facilities. 3.3.1. Human resources There was little difference in the availability of SBAs in health facilities that did and did not charge user fees. Almost all hospitals, and 53% and 56% of health centers that did and did not charge user fees had SBAs. However, the average ratio of SBAs to 1000 institutional deliveries did differ significantly: 60 per 1000 compared with 47 per 1000 for hospitals, and 50 per 1000 compared with 40 per 1000 for health centers that did and did not charge user fees, respectively. 3.3.2. Quality of service There was little difference in terms of quality of care provided between the health facilities that did and did not charge user fees. Just over half (58%) of hospitals that charged user fees provided the full package of EmONC compared with 53% of those that did not. Similar findings were seen for all the standard parameters of quality used in the emergency obstetric care assessment (Table 4): use of a partograph to monitor labor, availability of oxytocin, managing direct obstetric complications, and the availability of newborn resuscitation.
The case fatality rate (CFR) for direct obstetric complications varied significantly at 3.8% for hospitals that charged user fees compared with 7.1% for those that did not, and 1.2% for the health centers that charged user fees and 0.4% for those that did not. These contradictory findings require caution as they do not consider the condition of the patients going to the free facilities, the extent of underreporting of complications or deaths, and high rates of referral out, for example. The findings may also be due to sampling errors since there were very few severe obstetric complications managed at health centers. 4. Discussion This assessment of maternity services in Ethiopia has some limitations, which are described in detail elsewhere [12]. Facility records of deliveries, obstetric complications, cesarean deliveries, and deaths were often incomplete. Deaths, in particular, can be omitted or classified inaccurately, in part because maternal deaths from indirect causes are less likely to be found in the maternity or gynecology wards. Under-recording of complications (and deaths) will affect the direct obstetric case fatality rate. Assessment of equipment, supplies and drugs was encouraged but given the very long lists of these articles, not all items were observed. Although designed as a census of hospitals and health centers, ensuring that all hospitals, health centers, and higher clinics were visited proved challenging. The list of facilities registered in the Ministry database was incomplete. The coverage of facilities was incomplete in some regions. For security reasons, only 11 of the 26 facilities were visited in Somali Region, although most hospitals were visited. In Addis Ababa, the original list of facilities obtained from the local Regional Health Bureau was restricted to those that were licensed at that time. After the conclusion of data collection it was learned that some private clinics and hospitals were not included. Also, in Addis Ababa, some facilities were not forthcoming with data for some of the modules, and these data were not included in the analysis. Thus, it is likely that the indicators that reflect availability and utilization are conservative estimates.
314
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
The national emergency obstetric care assessment clearly shows severe underutilization of maternal health services (6.6% of the total expected annual births) and there are many reasons for this, including geographic distance, poor roads, lack of transport, perceptions of poor quality and availability of care, and economic constraints. A 2005 community- based survey by the safe motherhood program of Ethiopia found that decision-making about where childbirth takes place involves many members of the family and community, who retain control of the process and the outcome of births. Timely care seeking depends on the knowledge, understanding, and financial means of the household—usually controlled by the husband. Distance, cost, and considerable support for the cultural practices around birth hinder the use of modern health services [13]. Still, overall poverty seems to be the biggest reason for people not to seek care when sick. The average income in Ethiopia was US $0.68 per day in 2010 leaving little flexibility to manage unanticipated expenses. Other factors in addition to costs are physical access including referral and perceived quality of care. These should also be addressed as the removal of user fees is implemented [14]. Since there is already a policy in Ethiopia that maternity services (among others) at health centers should be free of charge, consideration needs to be given to regulation and enforcement. The waiver system is used to provide free services to the poorest at all levels, including hospital. Poorer households are preidentified to receive all health services free and are given certificates valid for 3 years. A recent review in Amhara region found that the revised fee waiver system had improved timely access for the poor to health services, but reimbursement to the facilities for waiver patients by the district offices varied from 25% to 100%. Some health centers received only US $0.40 per patient, which is much lower than the costs of a normal delivery. Problems were also identified with the fee waiver system such as targeting errors, limits to numbers of beneficiaries, and incomplete exemptions [15]. The recent National Health Accounts (NHA) IV Household and Utilization survey discovered that the poorest quintile of the population actually has the lowest percentage of fee waiver beneficiaries (data not shown) [8]. This provides evidence that the poorest are likely being missed in waiver targeting or facing other barriers to access of available services. The waiver system currently in operation in Ethiopia seems to be operating in the majority of government facilities, but there remain significant issues in terms of targeting the waivers to the neediest. More efficient systems of community-based targeting could be piloted during the roll out of the social and community-based health insurance systems in the country. The NHA also revealed that the national health expenditure (NHE) increased from US $522 million in 2004/05 to over US $1.2 billion in 2007/08. Per capita NHE also grew substantially, more than doubling from US $7.14 per capita per annum in 2004/05 to US $16.09 in 2007/08. This figure, however, includes funding received from all sources, including patients’ out- of-pocket expenditures. The major funding sources for reproductive health care in particular were international donors (36%), regional and local governments (28%), and households (25%) [8]. According to the Demographic Health Survey 2011 [5], the national average of institutional births is 10%: 51% for urban and 4% for rural areas. It is likely that the majority of users of maternity services have greater capacity to pay and are more motivated to seek care and therefore less influenced by the price. This is reflected in the findings that facilities that charge user fees see more patients on average than those that do not. The exemption policy regarding maternal services in place at primary healthcare unit level removes one of the key barriers to appropriate care seeking; if it is implemented correctly. A recent study in Ghana showed a nonstatistically significant decline in deliveryrelated deaths following the implementation of the exemption policy. Nor was there evidence that the severity of delivery-related obstetric conditions was significantly less before and after the exemption. In
contrast, interventions to improve access to care at the Juaben Teaching Health Center in the Ashanti Region of Ghana for example, led to a 3fold increase among women with complications seeking care and a 67% drop in referrals for treatment [16]. Our study showed that free maternity care is implemented in only a minority of facilities, with 65% of government health centers charging for services, supplies, drugs, or everything. Unlike waivers, there is no formal system of reimbursement of exempted services and health centers are expected to absorb these costs. At the same time, health centers are encouraged to raise and retain their own revenue through user fees for nonexempt services allowing them to improve their facilities. Some services on the exempted list are actually provided free of charge at the point of use (malaria treatment and immunizations, for example) and are those that are funded by international donors. Although the quality of many service parameters varies between facilities that charge and those that do not, it is clear that hospitals that charge are better equipped. One critical difference is the more favorable ratio of SBAs in facilities that charge compared with those that do not. This affects quality of care (for example as reflected by case fatality rates) and patients’ perceptions of quality of care (as reflected by utilization rates). The mandate to provide free maternity services in health centers to all women and the waiver system for hospital care for the poorest deprives health facilities of a significant source of revenue. In addition, the level of reimbursement of waiver services is not standardized and is insufficient to cover the costs of care. Providing emergency obstetric care at no cost to poor women is an important step forward in reducing the number of women who die in childbirth. However, the economics is complex. If care providers are not compensated in some way for this loss of revenue the quality of care provided will inevitably suffer. This is no small challenge. Health economists have developed such methods as low cost insurance schemes, increasing fees charged to patients better able to pay, installment payments, and voucher schemes. Governments must consider alternative sources of revenue to providers when considering mandates to provide free care to a part of the population. 5. Recommendations Generally speaking, when user fees cannot be removed completely, a targeted fee waiver for the poor and an exemption system for women suffering from obstetric complications, particularly those unable to pay, should be considered. Increasing coverage of health insurance can also increase availability of affordable maternity services. These are important steps to increase use of lifesaving services and to achieve MDGs 4 and 5. When progressive policies such as the fee waiver and exemption system are in place, consistent implementation is needed to ensure that the poorest are adequately targeted. Either all exempted services should be reimbursable to health providers or other means of cost recovery should be developed. Balancing these measures with maintaining high quality of care should not be overlooked. Addressing other barriers to access such as opportunity costs and distance remain a challenge. To enable free maternity services, health facilities should be adequately reimbursed for the essential supplies, commodities, and services rendered. The user fees can be compensated in different ways such as providing clean delivery kits, vouchers, or direct cash support based on the services provided. Communities should be informed of the eligibility criteria for fee waivers and services exempted from user fees. This would create demand and improve uptake of available services. For Ethiopia specifically, there is room to increase the government's allocation to the health sector. The Growth and Transformation Plan (2010–2015) of the Government of Ethiopia plans to double the government allocation to the health sector, which is a great opportunity to increase investment in the health system and maternal and newborn health in particular. Since waivers are issued
L. Pearson et al. / International Journal of Gynecology and Obstetrics 115 (2011) 310–315
and exempt services are reimbursed by Woredas (districts), there should be sufficient budget allocation from the central and regional governments for these activities. In the short term, financial support from development partners is important to bridge the financial gap. Acknowledgments The authors are grateful to the Ethiopian Federal Ministry of Health and Regional Health Bureau teams for their support. We recognize the effort of the many individuals at UNICEF, UNFPA, WHO, BETA, and AMDD who contributed to this Assessment. The assessment was jointly funded by UNICEF, UNFPA, and WHO. AMDD provided technical assistance. Conflict of interest None. References [1] Van Doorslaer E, Wagstaff A. Equity in the finance of health care: Methods and findings. In: Van Doorslaer E, Wagstaff A, Rutten F, editors. Equity in the finance and delivery of health care: An international perspective. New York: Oxford University Press; 1993. [2] Ridde V, Morestin F. A scoping review of the literature on the abolition of user fees in health care services in Africa. Health Policy Plan 2011;26(1):1–11. [3] Witter S, Mbengue D, Moreira I, De Brouwere V. The national free delivery and caesarean policy in Senegal: Evaluating process and outcomes. Health Policy Plan 2010;25(5):384–92.
315
[4] Richard F, Witter S, de Brouwere V. Innovative approaches to reducing financial barriers to obstetric care in low-income countries. Am J Public Health 2010;100(10): 1845–52. [5] Central Statistical Agency. Ethiopia Demographic and Health Survey 2011 (Preliminary report). Addis Ababa, Ethiopia: Central Statistical Agency; 2011. [6] Barnett I, Tefera B. Poor households’ experiences and perceptions of user fees for healthcare: a mixed method study from Ethiopia. Young Lives International Study of Childhood Poverty. Working paper no. 59. June 2010. Available at: http:// www.younglives.org.uk/our-publications/working-papers/poor-households2019experiences-and-perception-of-user-fees-for-healthcare. [7] 2nd Oak Foundation and Young Lives study consultation on vulnerable children, 6th December 2010, Addis Ababa. Unpublished. [8] Federal Ministry of Health. National Health Accounts IV: Household Health Service Utilisation and Expenditure Survey report; 2010. Addis Ababa, Ethiopia. [9] Federal Ministry of Health. Implementation manual for health care financing reports; 2010. Addis Ababa, Ethiopia. [10] Kruk ME. Bypassing primary care facilities for childbirth: a population-based study in rural Tanzania. Health Policy Plan 2009;24(4):279–88. [11] Kumbi S, Debebe A, Assefa G. Reasons for underutilization of emergency obstetric series in three Woredas of West Shoa Zone, Oromia region, Ethiopia: a qualitative study. Paper presented at the 12th Annual Conference of the Ethiopian Society of Obstetricians and Gynecologists; 2002. [12] Federal Ministry of Health. National Baseline Assessment for Emergency Obstetric and Newborn Care; 2008. Addis Ababa, Ethiopia. [13] Worren C. Care seeking for maternal health: challenges remain for poor women. Ethiop J Health Dev 2010;24:87–129. [14] Kruk ME, Mbaruku G, Rockers PC, Galea S. User fee exemptions are not enough: out-of-pocket payments for 'free' delivery services in rural Tanzania. Trop Med Int Health 2008;13(12):1442–51. [15] Save the Children UK. Situation analysis of health care financing reform. Implementation in five districts of south Wollo zone of Amhara region; 2009. Ethiopia: Successes and challenges. London: Save the Children. [16] Bosu W, Bell JS, Armar-Klemesu M, Tornui JA. Effect of delivery care user fee exemption policy on institutional maternal deaths in the central and volta regions of Ghana. Ghana Med J 2007;41(3):118–24.