Final Report Summary - FEMHEALTH (Assessing the impact of fee exemption on maternal health in West Africa and Morocco: new tools, new knowledge)
Executive Summary:
This report summarises some of the main conclusions and lessons from multi-disciplinary evaluations of four national policies to increase access to obstetric care in Burkina Faso, Benin, Mali and Morocco. These countries have all introduced universal subsidy or exemption policies targeting deliveries and caesareans (in some cases) in the last decade. The FEMHealth project used a variety of research tools to understand the origins, financing, implementation and effects of these policies, focussing not only on utilisation, health gains and financial protection but also on impacts on the wider health system and any unintended effects.
There is a growing consensus that maternal health outcomes can only be improved through policies and programmes that combine interventions to address the different causes of ill health and target multiple groups. Such policies and programmes are complex in nature as they involve coordination between different tiers in the health system and multiple actors including communities, health workers and managers. User fee exemption for delivery and emergency obstetric care (EmONC) is one such policy that has been introduced by many countries in Africa and globally with the aim of improving access to care and thus improving maternal and neonatal outcomes. However, the current evidence base regarding the impact of this policy is not well developed, in part because of evaluation designs that are not able to capture all the necessary information for policy-makers to make informed decisions. The FEMHealth project (2011-14) aimed to reduce this gap by developing research methodologies and tools that would lead to enhanced research on policy implementation, stronger evidence and improved dissemination.
The findings presented in this report are taken from 14 main research tools, most of which were used in all four study countries. They used mixed methods and included: document review; interviews with key informants at international, national and district level; analysis of routine financial and health information, as well as of secondary survey data; structured extraction from medical files; surveys of patients and staff; in-depth interviews with patients and observation of care processes. Conceptual frameworks linking all of these components were developed at the start of the programme and helped to integrate results.
Within each country, 4-6 study sites were chosen using criteria relating to minimum samples, but also a range of contextual criteria including poverty rates, utilisation levels, and distribution of population - the aim being to include a variety of contexts, including areas with higher barriers of poverty and low utilisation as well as ones where access was better prior to the policy introduction.
In order to study effects, FEMHealth used previous studies, where available, and time trend analysis. We also used cross-sectional analysis to understand how differences in outcomes might link to differences in implementation of the policies. Our overall approach was to construct a plausible narrative of policy introduction, implementation and effects, based on a triangulation of different sources and methods, and to observe any patterns across the different sites and contexts.
Key findings were as follows:
1. All three countries (recent data was lacking for Mali) have seen increasing access over time, and while the policies may have contributed to their continuation into the current period, they have not apparently accelerated that trend
2. Analysis of changing utilisation by socio-economic group shows a narrowing of inequalities for all three countries with survey data. Again, this is part of a longer term trend and relates in part to the prior high supervised delivery rates of higher income quintiles. It is not possible to quantify the role of the policies, though they are likely to have contributed to some extent
3. The overall evidence suggests a significant reduction in household payments for the targeted services, ranging from 60 to 90%, depending on the delivery type and country
4. However, even in relation to the package of care which was supposed to be covered, households continued to pay sums which amounted to a varying proportion of their overall expenditure (small in Morocco, intermediate in Burkina and Benin and substantial in Mali)
5. Cross-sectional analysis shows that quality of care is variable across sites and that there is no evidence that those hospitals which are implementing the policy effectively are providing worse care. In some cases, such as the Burkina Faso sites, it is the opposite pattern
6. There were no systematic effects on the wider health system or on untargeted services
7. From a financial perspective, the financial burden of the policies is manageable. They cost 2.5-3.5% of public health expenditure in 2011, and were funded from national resources
Some notable strengths were found across the countries. There is considerable political commitment to these policies and to maintaining the gains they have generated. They have been relatively thoroughly implemented, without the budget shortfalls which have been documented in other settings. They have also, in some cases, like Morocco’s, been accompanied by the additional supply-side improvements which are required. Our research also found an underlying support for them, not only from beneficiaries but also from key actors within the health system (health district managers, hospital management teams, specialists, nurses and midwives). The policies were generally considered to be relevant and important. In some cases, they also appeared to generate efficiency savings.
However, the heavy emphasis of these policies on caesareans (in two out of four countries) has been problematic in a number of ways: caesareans can save lives, but even if utilisation increases it is not easy to know if the right (medically indicated) women have received care. Moreover, the use of caesareans is heavily skewed to the rich and to urban areas, meaning that the benefits of the funding will almost automatically be biased in favour of the rich. It is an intervention which in some contexts needs boosting, but in other contexts (or for some groups) needs controlling. It can be induced by suppliers and patients for the wrong reasons, and carries medical risks.
Moreover, in other countries, where fixed payments to facilities were high and where onward charges to women continued illegally under the policy, the main winners were the hospitals, which gained financially from the policy. Unfortunately there was limited evidence of this being used to improve the overall quality of care or to strengthen the health system as a whole.
We have found a range of outcomes in different contexts (positive and negative across different sites within the same country). This underlines the importance not just of policy design but also context and the institutional and organisational frameworks into which policies are introduced.
The policies are relatively recent and it is early to make a final judgement on this question, but in general, they are likely to have played a part in supporting continued improvements, along with other changes and investments. In the case of Morocco, for example, there were a number of parallel investments within the overall Action Plan and many households prior to the policy benefited from a card providing exemption from all payments for low income households. In this context it would not be expected that the free delivery care at hospitals would lead to a dramatic shift in behaviour. Rather, it should be seen as part of a continuum of measures to increase facility deliveries over time.
The policies are universal in design and should benefit all women. However, non-financial barriers are more significant for women in rural areas, particularly in relation to transport. Moreover, the policies support those who use the services, which in all countries were skewed towards the better off households before the policies (especially for caesareans).
Changing that means changing care-seeking behaviour at delivery, which as our research has shown is quite a difficult task, especially over the short term. This requires raising awareness of the policies, especially amongst non-users and more remote women, improved physical access and reassuring women in relation to their reception, the costs they will face and the support they will receive during deliveries.
Evidence suggests that all of these require more effective actions in the study countries. A significant proportion (0-35%, depending on site) was unable to pay, even after policy implementation. Moreover, women reported a lack of certainty about what they should pay or not which not only increases financial problems but also clouds the relationship with providers. This indicates that there is plenty of scope to increase the financial protection offered by the policies. For example, in Burkina Faso, the payment of the residual 20% for indigents remains to be implemented.
The case studies found that the targeted policies have not created opportunities to strengthen the stewardship function, or at least, these opportunities have not been taken. This may present a challenge beyond the scope of a single national policy.
Project Context and Objectives:
PROJECT CONTEXT AND OBJECTIVES
There is a growing consensus that maternal health outcomes can only be improved through policies and programmes that combine interventions to address the different causes of ill health and target multiple groups. Such policies and programmes are complex in nature as they involve coordination between different tiers in the health system and multiple actors including communities, health workers and managers. User fee exemption for delivery and emergency obstetric care (EmONC) is one such policy that has been introduced by several African countries with the aim of improving access to care and thus improving maternal and neonatal outcomes. However, the current evidence base regarding the impact of this policy is not well developed, in part because of evaluation designs that are not able to capture all the necessary information for policy-makers to make informed decisions. The FEMHealth project (2011-14) aimed to reduce this gap by developing research methodologies and tools that would lead to enhanced research on policy implementation, stronger evidence and improved dissemination.
The overall aims of the project were: (1) to develop new methodological approaches for the evaluation of complex interventions in low income countries, (2) to improve the health of mothers and their newborns by performing comprehensive evaluations of the impact, cost and effectiveness of the removal of user fees for delivery care on maternal and neonatal health outcomes and service quality, and (3) to improve the communication of this evidence to policy-makers and other stakeholders. This report summarises evidence in relation to the second objective, drawing across a range of research components and tools.
At the start of the project, a set of core research questions were developed which outlined the main levels, domains and topics which FEMHealth would seek to investigate (Figure 1). These form the broad structure of this report, which flows from an analysis of the drivers behind the policies, their objectives and formulation at the national level to analysis of how they interacted with and impacted on district-level health systems, and, finally, their effects and effectiveness at community and household level.
RESEARCH METHODS
At the start of the project, a set of core research questions were developed which outlined the main levels, domains and topics which FEMHealth would seek to investigate. The first set (within work package 2) relate to health policy and health financing analysis. The second (WP3) focussed on analysis of the health district implementation and effects. The final work package (WP4) analysed changed to utilisation and quality of care. These form the broad structure of this report, which flows from an analysis of the drivers behind the policies, their objectives and formulation at the national level to analysis of how they interacted with and impacted on district-level health systems, and, finally, their effects and effectiveness at community and household level.
The findings presented in this report are taken from 14 main research tools, most of which were used in all four study countries. They used mixed methods and included: document review; interviews with key informants at international, national and district level; analysis of routine financial and health information, as well as of secondary survey data; structured extraction from medical files; surveys of patients and staff; in-depth interviews with patients and observation of care processes. Conceptual frameworks linking all of these components were developed at the start of the programme and have been described elsewhere (Marchal et al. 2013).
Within each country, 4-6 study sites were chosen using criteria relating to minimum samples, but also a range of contextual criteria including poverty rates, utilisation levels, and distribution of population - the aim being to include a variety of contexts, including areas with higher barriers of poverty and low utilisation as well as ones where access was better prior to the policy introduction.
More details on the methods can be found in the constituent reports, which will be made available on the FEMHealth website (www.abdn.ac.uk/femhealth).
All protocols received in-country ethical approval – in October 2011 in Burkina Faso, January 2012 in Morocco, March 2012 in Benin, and July 2012 in Mali. In addition to the approval from the ethics committee, administrative authorization was requested and obtained from the regional health departments, districts, hospitals and at national level in all four countries. The component tools supported by thematic work packages were also approved by the ethics committees at the LSHTM and ITM.
The main limitation faced by the study as a whole was that in each case, policies had been introduced nationwide and some years earlier, with no control areas and no formal baseline data. In order to study effects, FEMHealth used previous studies, where available, and time trend analysis. We also used cross-sectional analysis to understand how differences in outcomes might link to differences in implementation of the policies. Our overall approach was to construct a plausible narrative of policy introduction, implementation and effects, based on a triangulation of different sources and methods.
In the case of Mali, civil conflict has meant that the full data set could not be collected. Consequently, this report contains some results for Mali, but not the full evaluation results which were obtained from the other three countries.
Project Results:
CONTEXT AND DRIVERS BEHIND POLICIES
One of the FEMHealth research strands aimed to understand the origins of the policies in this region – why had so many countries adopted similar policies over a short period? What were the drivers behind this and how far had they been influenced by one another and by international actors?
One set of drivers related to context and the recognition by decision-makers that socio-economic factors were behind low overall supervised delivery rates in Mali, Burkina Faso and Morocco (Benin already had relatively high facility deliveries at 80%) and large inequalities in all four countries. Design of policy varied however: Morocco took the most wide-reaching approach, embedding the free obstetric care at hospital level within a wider action plan which included investments in quality of care and improved governance. Mali and Benin were the most specific, focussing purely on caesarean sections and some related complications. Affordability was one factor behind the narrowing down of services in Benin.
The hypothesis that international actors might have been behind the proliferation of related policies in the West Africa region was not sustained: interviews suggested that while international influences have been important in shaping the global climate which permeate the four case study countries, the decision-making and elaboration of the policies were dominated by local factors. International actors may have lost some credibility through changes in stance on issues like user fees, with strong but changing messages over the past two decades.
Countries which were once the heartland of the Bamako Initiative have been amongst the most active in taking up selective exemptions (which were seen as more acceptable and affordable than broader approaches to fee removal). The emphasis on the MDGs and, now, on universal health coverage have influenced the underlying discussions, and these free care policies have to be seen in a context of proliferating exemptions in the countries and region for many different vulnerable groups and priority services, though variably implemented. Shame at performing less well in relation to neighbours was highlighted as a driving factor for Morocco. Personal political leadership was seen as critical in all contexts to enabling the policies to be realised – particularly for setting out a vision, and mobilising funds and support. Evidence, while it was marshalled quite thoroughly in two countries (Morocco and Burkina Faso), was used more to assist with planning implementation details than in propelling the original policy adoption itself.
In relation to learning across countries, there was some evidence that countries had looked with interest at the experience of their neighbours who had started implementing earlier, but no formal channels of transmission of detailed lessons were identified.
DESIGN AND IMPLEMENTATION OF THE POLICIES
While there are many common elements, the policies differed in some respects. The government in Burkina Faso opted for a subsidy policy supporting 60 to 80% of the costs of obstetric and neonatal care in all public health facilities in the country. This initiative differs from what is in place in Morocco where normal delivery and caesarean sections are totally free at all levels of public hospitals. The policies in Benin and Mali are also different from the Burkinabe one since they are based on 100% reimbursements of caesarean costs alone. In the case of Benin, this covered accredited confessional and private facilities as well as the public sector.
All medical costs associated with the target services within hospitals (and health centres in Burkina) were included in principle in the package. None of the countries covered transport to the first level facility, but all claimed to cover onward referral transport, though in practice patients often paid. For Morocco, food within the hospital was covered, but this was not the case for the other countries.
Implementation arrangements also varied: most policies were managed by a national committee but in the case of Benin, a dedicated autonomous agency was established to manage the policy. These differences are reflected in some of the research findings: for example, the centralised model adopted by Benin may explain why there were no discrepancies or major delays in reimbursement flows to facilities. However, there were also downsides, in terms of a lack of involvement of the zone in managing and monitoring the policy (the chain of command went straight from the agency to the hospitals, without involving district managers).
FINANCING
In all four countries, the policy is financed almost entirely by the state, with a notable absence of direct donor financial support and indirect donor support for the sector ranging from 0.5% of public health spending in Morocco to 36% in Burkina Faso.
The three sub-Saharan African countries reimburse retrospectively according to the number of services provided, using fixed payments per caesarean in Mali and Benin, and reimbursement of actual variable costs in Burkina Faso. For Morocco, hospital budgets have been boosted to reflect rough estimates of revenue likely to be lost due to the withdrawal of fees for deliveries, but these bear no relation to actual workload and have not increased over time (indeed all subsidies for hospitals were unpaid in 2011). Additional investments in staffing and improved access and care were made in Morocco, which was not the case for any of the other countries. These features are reflected in some of the findings on implementation and effectiveness: quality of care scores, for example, are relatively good in Morocco.
The payment systems also create specific incentives. For Burkina Faso, the payment of actual costs means that there is no incentive to over-provide or to skimp on quality, but on the other hand, there is no surplus to reinvest (the costs are just the variable ones, such as drugs and supplies, so there is no contribution to general facility running costs). In addition, the workload implied by billing by item is considerable, which is why this factor caused discontent amongst the staff in Burkina Faso, more than in any other country. By contrast, the fixed tariff in Benin was found to overpay caesareans at all levels of hospital (relative to actual production costs but also relative to previous payments from users). On the positive side, examples of managers reinvesting the surplus in improving overall services were found. On the negative side, there was a perception that staff were sometimes too eager to do a caesarean, rather than letting women try for a normal delivery. In addition, despite the generous payments, some hospitals continued to make women pay for items which should be free (see below).
The scale of investment is hugely different across the countries, with Morocco spending in the region of 24.5 million Euros on its overall action plan in 2011, compared to 3.2 million Euros in Benin, and 415,000 Euros in Burkina Faso. Per delivery covered that equates to 1.3 Euros for all deliveries (Burkina), 152 Euros per caesarean (Benin), and 797 Euros for all deliveries combined (Morocco). Clearly, these differing scales of investment should shape our expectation of results.
As a proportion of public health expenditure in 2011, the policies absorbed around 2.5% in Morocco, 3% in Benin and 3.5% in Burkina Faso – not insignificant but all potentially sustainable, if the policies are seen as effective.
HEALTH SYSTEMS EFFECTS
Human resources
The only country to accompany its policy with a significant increase in staff was Morocco, which increased the deployment of midwives and hospital specialists as part of its overall action plan. For other countries, there was no evidence of increased numbers of staff in any large measure, though in Benin, the additional resources provided by the policy allowed some local changes to improve staffing. Reported working hours for the main groups of staff concerned with obstetric care remain reasonable even after the policy, across all four countries, but workload (patients seen and deliveries attended) are reported to have increased for three out of four of the countries (Mali being the exception). In Burkina Faso, the administrative workload imposed by the policy was a particular concern. The consensus was that the policies have not affected their remuneration (as there are no direct financial incentives linked to it for any of the staff in any of the countries).
Across all countries the majority of staff surveyed feels positive about the wider effects of the policy, believing that the policy has increased access to supervised deliveries, has benefited the poor, and has improved the quality of care, including through improvements to drugs and supplies. This also impacts positively on health worker working conditions and satisfaction in some countries, like Burkina Faso and Mali.
Facility finances
In Morocco, despite the increased budget to accompany the implementation of the free caesarean and delivery policy, the general revenues of the health facilities have not been affected positively. In 2011 and during the first half of 2012, no financial resources were provided at all. This has affected the financial reserves of the health facilities that are still implementing the gratuity policy. For Benin, the costing revealed that the current tariff is beneficial to hospitals, paying considerably more than it costs to provide the service, though the surplus varies by type of hospital. In addition, payments have been regular. For Mali, the calculus is different, with the reimbursements not fully covering costs of provision. In Burkina Faso, payments are often delayed and do not include any operating costs beyond the direct inputs needed for the service.
It might be expected that extra costs would be levied from users in countries where facilities are poorly rewarded for providing care, however analysis of household payments does not support this. In absolute terms, women reported paying the largest amount overall for caesareans in Benin (60 Euros on average), where the payment for the policy is most generous. This suggests that organisational culture and other factors play a role in the levying of additional payments from users, rather than actual financial need from the facility perspective.
IT systems
There has been no broad effect of the policies on health information systems, either positive or negative. In Benin, there has been a policy-related improvement in the obstetric information gathering, but this has not cascaded into other areas of health information. In Burkina Faso, a specific system of collecting information on the policy was put into place, and it is operating with variable success. As for Morocco, the wider information system is unchanged and is generally seen as onerous for health staff.
Drugs and supplies
All four countries used kits to support the implementation of the policy, with varying results. The policy in Morocco was accompanied by a large increase in kits, which meant that drug supply improved, though in some places, kit numbers were well in excess of need. For Burkina Faso, drug supply also improved, but management of kits sometimes lacked transparency and stock-outs did sometimes occur. In both Morocco and Burkina Faso, equipment was felt to be lacking, with no provision made to improve working conditions. In Benin, the kit contents were perceived to be too generous at first, leading to misuse, and then too restrictive, leading to costs to users for items not included in the kits. In general, it could be said that attention was paid to providing specific drugs and items needed for obstetric care, but there was no wider investment in the supply system and a continuing focus on the use of kits, which as seen, may not be the most efficient method of organising supplies.
Management
The policies introduced threats and opportunities for local managers, and examples were found of positive and negative adaptation. Generally, managers had not been much involved in the development of the policy, and in many cases detailed guidance on how to implement it was lacking. Management teams have leeway to interpret the policy. In some cases, it was found that actors such as directors and specialists used their power position to adapt the policy to their own benefit, e.g. in some study sites of Benin or Morocco, patients were still charged fees, because staff were compensating for the free caesareans by charging something else to the patient. These effects can be moderated or avoided by capable management teams and adequate supervision by programme managers, as seen in other study sites of the same countries, where district health managers and or hospital directors are positively engaged with the policy implementation and the protection of service users.
In most study sites in Benin, Mali and Burkina Faso, the management teams did not have a large absorption capacity to take up the new tasks without additional resources. When the reimbursements were late or inadequate, the implementation halted. In Morocco, this was not the case, as the capacity of the teams, the existing implementation infrastructure in terms of human resources, facilities, and equipment was adequate to take up the new patients.
FACTORS BEHIND DIFFERENTIAL IMPLEMENTATION
The analysis of divergent responses of different cadres within our realist evaluation case studies indicated that the adoption of the policy is explained in part by the configuration of autonomy, decision space and motivation of these actors and by organisational, institutional factors and contextual factors. Different responses were found by teams in different sites within the same country, indicating that beyond policy design and national features, there is an inter-play of local factors which influences whether the policy is blocked, adopted or adapted. This highlights the importance of reinforcing the stewardship function, not only in relation to these policies but more generally. The role of ensuring public accountability, in particular, was found to be underdeveloped or even not mandated clearly. In any case, lack of effective stewardship allowed faulty implementation processes to continue in many of the sites.
IMPACT ON UTILISATION
• For Burkina Faso, there is a statistically significant decline in the rate of increase facility deliveries (i.e. between 2002 and 2007 there is a 12% relative increase each year, whereas between 2007 and 2010 there was a 6% relative annual increase) - so whilst the overall rate is increasing in absolute terms, it is doing so less rapidly post-2007. For caesareans, there is no evidence of a change in either direction.
• For Morocco, there is no evidence of a change post-2008 in facilities deliveries: the gradient is flat. There is actually a significant decrease in caesarean rates after 2008, though this may be related to the 2008 data point being erroneously high. In either case, there is no evidence of a change in gradient.
• For Benin, there is an upward trend but no significant change in gradient post-policy
• For Mali, we have no comparable post-policy data, though the change may be modelled when the next DHS dataset becomes available.
It is evident overall that countries have made progress over the past 15-20 years, and these policies may have contributed, but there is no evidence of that as yet. It is almost certainly too early to tell, as we have only 2-3 post-policy data points in each country, and the varying implementation documented by FEMHealth also underlines the need to be cautious about assuming immediate effectiveness of policies.
Furthermore, the Morocco caesarean section rate is at a level where any increase (even if true) would be unlikely to lead to further reductions in maternal mortality, and if anything the concern is excess and unnecessary caesareans among certain groups.
IMPACT ON OTHER (UNTARGETED) SERVICES
The research looked in a systematic way for the impact of the policies on untargeted services, in order to capture any unintended positive or negative effects, but found no major effects. Some positive effects of additional resources introduced by the policies were documented (e.g. in Morocco and Benin), as well as from wider utilisation uptake in Burkina Faso, linked to the policy. On the negative side, some examples were found where the policy encouraged resources (such as staff) to move from untargeted to targeted services. More significantly, in Benin, there was some evidence of supply-induced demand for caesareans, at the expense of normal deliveries. Trend analysis of provision of general medicine and paediatric services in all countries and sites did not reveal any evidence of distortions linked to the exemption policies.
IMPACT ON QUALITY OF CARE
Overall, women’s perceptions of the quality of the services were generally high or very high, and did not correlate well with technical quality of care scores (being highest in Burkina Faso, where technical scores were lowest). The quality of neonatal care, measured by the number of omissions in routine neonatal procedures, was very poor in some hospitals in Benin and Burkina Faso, and generally poorer than the quality of maternity care. Median delays in receiving caesarean sections were above the expected threshold of 1 hour in most hospitals (except in Morocco) and were the highest in Benin hospitals where the policy was designed to facilitate access to life saving emergency surgery. Hospitals in Morocco performed consistently better than hospitals in the other countries.
In observations and interviews in Benin and Morocco, incidents of poor communication were observed, including lack of informed consent for surgical care and poor bedside manners. In Benin, while poor interpersonal relationships in maternity wards had been documented in previous studies, in this project these were attributed by several respondents to perverse effects of the free caesarean policy.
Hospitals are still receiving many cases of near-miss, particularly maternal near-miss, with hospital incidence ranging from less than 2% in Morocco to 14% in Benin. As cases of maternal near-miss are women who nearly died and were saved in extremis, there is still a lot of progress to be made in the organization of the health services in order to reduce the burden of several morbidity and mortality in the focus countries.
Our key hypotheses included that hospitals/districts with lower user fees cost may register shorter delays and fewer adverse events because women may arrive earlier in facilities; but that on the other hand, an increase in volume of patients, if not met with an increase in human resources, might lead to a deterioration of the quality of maternity and neonatal care. While in Burkina Faso, there seemed to be a positive correlation between average omission scores, average delays and the success of the implementation of the policy (as measured by excess payments made by patients under the policy), limited relationships exist between the omission and implementation score in Benin, implying that quality of the care provided was affected by many factors which may be quite independent from policies designed to increased access.
IMPACT OF POLICY
Awareness of the policies was relatively low, ranging from 20% amongst women who had delivered in Benin to 53% in Mali (but much lower in some districts). As these interviews were conducted with women who had used services and had already delivered, wider awareness can be assumed to be even lower. Detailed knowledge of entitlements was very low. Clearly, if policies are to influence care-seeking a greater communication effort is needed, especially for poorer and more remote women (awareness tended to rise for higher quintiles).
Impact on delay in seeking care and health seeking behaviour
In general, median reported delays in leaving home, arriving at facilities and being seen were acceptable, although there was a large variation between sites and by type of delivery, and a wide range within the responses in general. The median delay in leaving the house was rather high in Benin. In addition, it is important to remember that those interviewed represent those who were able to access care, and these are often not the most disadvantaged women. Perceptions of need, the availability of transport and the availability of the key decision-maker (who varied by site) emerge as significant.
In-depth interviews with users suggest an appreciation of the policy and that the policy did address some of the key barriers to access. However, it did not necessarily change health seeking behaviour. In Morocco, interviews showed that the choice of location for deliveries was made largely according to expected comfort, care and monitoring (for example, at home), and reassessed in cases where outside help was decided to be necessary as a matter of urgency. Transport was a key barrier. At the hospital, the absence of a doctor, the gaps in surveillance, inadequate resources, and tedious negotiation process to receive the desired care were all aspects known and anticipated by women when considering their recourse to care.
In Benin, there was also no evidence from the interviews or observations that women modified their decision to seek out skilled professionals at birth due to the policy. While the policy was appreciated and the costs of caesarean sections were considered to be more affordable than in the past, the policy does not erase the fear of the caesarean section as a medical procedure and the threat of loss of life. Decisions about where to give birth were based on hospital reputation, convenience of access and past delivery experiences.
Impact on inequities of access
In all three countries for which there is recent household survey data (i.e. excluding Mali), the relative inequity between the poorest and the richest has declined over time (in that there have been bigger gains among the poorest). The policy may have contributed to this but this is a longer term trend, and one which followed to some extent from the fact that richer women already had high coverage. In all three countries there remains substantial inequity in utilisation of care.
Financial impact for households
In relation to the amounts which should have been paid by households under the policies, we found that they are paying excessive amounts in all countries, though the excess payments are relatively low for Morocco (2% of their total payment for caesareans and 6% for normal deliveries), which indicates a relatively effective implementation of the policy. By contrast, in Mali, 49% of the household payment is excessive. Intermediate proportions of 13% (for Benin) and 17% (for Burkina) were found for caesarean sections. The absolute amounts paid in Benin were higher than for Burkina, as the overall payments were higher.
Patterns of payment across quintiles and across rural and urban areas varied. Large proportions of the households in all settings and quintiles incurred catastrophic costs, even under the current policies (and a higher proportion in rural than urban areas in all four countries).
Certain costs, including transport, cost of companions, care of newborn (in some countries), tipping of health workers and supplementary drugs remain a burden. Lack of clarity on charging can also reduce predictability of costs and cause anger and confusion for women and their families.
Between 0 and 35% (depending on the site) of households surveyed had been unable to make the requisite payments, with those in lower quintiles more likely to report this than higher. In general households coped by using savings and getting help from family and friends, but some had to sell productive assets such as land to cover the bills. Very few had health insurance membership and in general, health insurance was not a protective mechanism for most households, as it is pro-rich in its distribution.
Looking at the difference between recorded payments prior to the policies and average payments now, households have made a substantial financial gain. In Burkina Faso, there was a reduction of 71% for deliveries of all kinds. In Morocco, the gain was lower for normal deliveries (62%), compared to 92% for caesareans. The estimated saving for caesareans in Benin was in the region of 74%, compared to 78% for Mali.
CONCLUSIONS
1. Did the policies increase access to obstetric care?
Analysis of secondary survey data, looking at trends over a longer time span, did not demonstrate a significant increase in utilisation associated with the start of the exemption or subsidy policies in any of the three countries for which we have data (with the exception of Mali). All three countries have seen increasing access over time, and while the policies may have contributed to their continuation into the current period, they have not apparently accelerated that trend.
The policies are relatively recent and it is early to make a final judgement on this question, but in general, they are likely to have played a part in supporting continued improvements, along with other changes and investments. In the case of Morocco, for example, there were a number of parallel investments within the overall Action Plan and in addition, many households prior to the policy benefited from a card providing exemption from all payments for low income households . An estimated 70% of the population were covered by these cards. Although the cards were not fully respected or used (Hodgkin, Krasovec K., El-Idrissi, Eckert, & Karim 2005), in this context it would not be expected that the free delivery care at hospitals would lead to a dramatic shift in behaviour. Rather, it should be seen as part of a continuum of measures to increase facility deliveries over time.
2. Which groups benefited most?
Analysis of changing utilisation by socio-economic group shows a narrowing of inequalities for all three countries with survey data. Again, this is part of a longer term trend and relates in part to the prior high supervised delivery rates of higher quintiles. Gains, where they occurred, were likely to be amongst those with lower use at the start. It is not possible to quantify the role of the policies, though they are likely to have contributed to some extent.
The policies are universal in design and should benefit all women. However, non-financial barriers are more significant for women in rural areas, particularly in relation to transport. Moreover, the policies support those who use the services, which in all countries were skewed towards the better off households before the policies came into effect, and even more so for caesareans. This inevitably means that a disproportionate part of the benefits are captured by better off households. A simple breakdown of our exit interview sample by quintile, across the countries, demonstrates this (Table 23). Relatively few of the users were in the poorest quintile. For policies focussed on caesareans, the three better of quintiles hold the bulk of users. For those with broader policies, there is more evidence of benefits for the poor (e.g. quintile 2).
Changing that means changing care-seeking behaviour at delivery, which as our research has shown is quite a difficult task, especially over the short term – it means raising awareness of the policies, especially amongst non-users and more remote women, ensuring physical access and also reassuring them in relation to their reception, the costs they will face and the support they will receive during deliveries. Evidence suggests that all of these require more effective actions in the study countries. In Burkina Faso, the payment of the residual 20% for indigents remains to be implemented.
3. Were they effective in reducing financial burdens?
The overall evidence suggests a significant reduction in household payments for the targeted services, ranging from 53% for caesareans in Benin to 92% for caesareans in Morocco. Some of the costs were expected to remain as the free care policies focus on in-facility costs. However, even in relation to the package of care which was supposed to be covered, households continued to pay sums which amounted to a small proportion of their overall expenditure in Morocco, intermediate in Burkina and Benin and the large majority in Mali. A significant proportion (0-35%, depending on site) was unable to pay. Moreover, women reported a lack of certainty about what they should pay or not which not only increases financial problems but also clouds the relationship with providers. This indicates that there is plenty of scope to increase the financial protection offered by the policies.
4. What is the overall evidence on effects on quality of care?
Quality of care is complex to measure and baseline data was not available to measure trends over the period (and even with baseline data, attribution to the policy would not have been possible, as quality is affected by so many factors). However, cross-sectional analysis has allowed us to conclude that the quality is variable across sites and that there is no evidence that those hospitals which are implementing the policy effectively are providing worse care. In some cases, such as the Burkina Faso sites, it is the opposite pattern. We can therefore conclude that the policy has not systematically weakened quality. Qualitative evidence has uncovered a large number of positive and negative effects. The realist evaluation case studies suggest that whatever the policy designs, there is scope at the local level to adopt the policies in a positive way. The focus should be on reinforcing competences and institutional arrangements to enable such positive management of resources.
5. Did the policy strengthen or weaken the local health system?
The POEM studies generally indicate that the policy did not reinforce any of the non-targeted services. Often, it led to improvement of resource availability for targeted services, changes in the health information system or competences of providers, but not beyond the sphere of maternal and child health, and specifically maternity and theatre. No changes were documented in general management or organisation of services. POEM and realist case studies of policy adoption have shown that there is no effective stewardship function in the majority of the sites, defined as coordination, management and regulation of all health actors and safeguarding the public interest. The targeted policies have not created opportunities to strengthen the stewardship function, or at least, these opportunities have not been taken. This may present a challenge beyond the scope of a single national policy.
6. What were the main strengths and weaknesses of the policy?
A detailed analysis of strengths and weaknesses are presented in the country reports. However, as a group, certain features stand out. On the positive side, and compared with the documented experience of other countries with similar exemption policies:
• The policies have been relatively thoroughly implemented: despite some gaps and lapses, the policies have been put into effect in a serious way
• They have not been affected by budget shortfalls, which undermined the effectiveness of similar policies in countries like Ghana (Witter & Adjei 2007)
• They have, in some cases, like Morocco’s, been accompanied by the additional supply-side improvements which are required to meet the additional demands
• There is an underlying support for them, and not only from beneficiaries: most actors within the health system (health district managers, hospital management teams, specialists, nurses and midwives) reacted positively to the policy in interviews. The policy was generally considered to be relevant and important
• They have achieved substantial reductions in household payments, which will over time contribute to poverty reduction and reduced inequalities of access
However, there are also some weaknesses. These include:
• A package of care which in some cases (Benin and Mali) will not address all of the main causes of maternal and neonatal morbidity and mortality, and whose impact on these can therefore only be expected to be modest, even if well implemented
• Poorly calibrated provider payments, for those using fixed payments, which either over-incentivise (in the case of Benin) or under-fund (in the case of Mali). Both of these result in perverse effects
• Lack of clear and well disseminated operating documents, which enable staff and clients to be clear about how the policy will work and what is covered by it
• Too limited attention to the quality of care offered by the facilities covered by the policy; for newborns in particular it has been found to be sub-standard
• Lack of involvement in most cases of managers, staff and communities in developing and monitoring the policy in order to increase ownership and control abuse
• No policy has completely reduced the officially exempted costs to zero; although magnitudes of unwarranted payments vary in scale, all countries need to more effectively regulate providers and stop illicit payments from patients
• By their very design, the policies are unable to address some of the main barriers faced by women, such as inability to physically access health care; additional actions are needed to ensure that benefits can be equitable
In addition to the weaknesses of the policies themselves, there are underlying systemic weaknesses which undermine policy effectiveness, such as lack of effective stewardship at the local level, drugs supply and distribution systems which are not reliable and poor provider-patient relationships in some areas.
7. Are these policies likely to be sustainable?
From a financial perspective, the financial burden of the policies is manageable. They cost 2.5-3.5% of public health expenditure in 2011, and were funded from national resources. There is considerable political commitment to maintaining the gains they have generated. The main challenge for sustainability will be developing national plans to move toward universal health coverage which can integrate the patch-work of different exemption and financial protection policies which are now multiplying in all of the study countries. In the course of this, the lessons learned through the FEMHealth evaluation should be borne in mind in order to maximise benefits and minimise risks.
8. Are they cost-effective overall?
The policies in Burkina Faso, Benin, Mali and Morocco were strong national initiatives which aimed to improve maternal health and to increase access to obstetric care. On one level, the evaluation produces inconclusive results: we observe positive trends in relation to supervised deliveries and caesarean sections and a narrowing of inequalities in all three countries for which recent data is available but cannot attribute these to the policies. There is no significant change in the trends which coincides with the introduction of the policies. It is likely that they have contributed to the ongoing trend, but this is only speculative.
The heavy emphasis of these policies on caesareans (in two out of four countries) has been problematic in a number of ways: caesareans can save lives, but even if utilisation increases it is not easy to know if the right (medically indicated) women have received care. Moreover, the use of caesareans is heavily skewed to the rich and to urban areas, meaning that the benefits of the funding will almost automatically be biased in favour of the rich. It is an intervention which in some contexts needs boosting, but in other contexts (or for some groups) needs controlling. It can be induced by suppliers and patients for the wrong reasons, and carries medical risks. While policies to reduce the costs of caesareans (which are high cost, potentially catastrophic events from a households perspective) can provide real financial benefits to households, as these have done, from a public health perspective a wider policy covering a range of life-threatening obstetric complications and also the pathway to them (facility deliveries) is preferable. Morocco and Burkina Faso illustrate this approach.
On that basis, should we conclude that the policies were not cost-effective, or did not represent value for money for the national budget? This requires a more nuanced judgement. One of the underlying objectives was to reduce the burden on households of this essential service, and so change behaviour. The evidence of the FEMHealth research suggests that while financial barriers are significant and are connected to many other barriers (physical, cultural etc.), on their own their reduction does not change behaviour, unless it is connected with a positive shift in other aspects, such as perception of quality and responsiveness. Policies on financial access therefore need to be designed with improvements to these other facets in mind, as an integral part of their design. Behaviour change also has to be measured over a longer period, as habits in relation to significant services such as delivery change slowly.
Looking at a simple comparison of the funds spent by government on the policies versus the estimated gains made by households gives another insight into the value for money question. In Burkina Faso, the average expenditure per delivery is lower than the average gain per household with a delivery. There is therefore a net gain, which probably reflects the payment system and the fact that facilities are providing care without fully recovering their costs. If they are able to do this and still provide adequate care without passing additional costs to women, then the policy is leveraging an efficiency gain in the health system. This is indicated by our results. For Mali, the net balance for caesareans is also positive, though some gaps in our data do not allow us to interpret fully how this has impacted on the health system. For Morocco, the unit cost information from the government side is missing.
For Benin, the calculus is different as the government is paying 100,000 FCFA per caesarean and our results suggest a household gain in terms of reduced costs of around the same magnitude. Societal gains might have been made if the policy had not set a fixed tariff which was too high for average production costs and then also failed to control additional payments which were levied by some hospitals on women and their families. Some of the winners in this scenario are the hospitals which have gained financially from the policy, but unfortunately there is limited evidence of this being used to improve the overall quality of care or strengthen the health system as a whole.
Wider impacts, positive and negative, intended and unintended, are also to be taken into account in coming to an overall judgement about these policies. We have found a range of these but they vary by context and suggest a variety of outcomes can be expected from these policies, depending on their features but also the context and the institutional and organisational frameworks into which they are introduced. This complexity means that no one simple answer to the overall evaluation can be produced for all settings.
The overall recommendations arising from the evaluations are closely linked to tackling the weaknesses outlined above, not only in relation to the policies but also the underlying systemic challenges.
Potential Impact:
Some of the key messages arising from the FEMHealth research are as follows:
1. A broad package of care should be exempted if the goal is improved maternal and neonatal health. Policies which focus on caesareans alone cannot address the main causes of death and disability
2. Exemption and subsidy policies can provide financial protection if well designed and implemented.
3. There are high risks of favouring better off households unless exemption policies are accompanied by concerted efforts to address other barriers (physical, cultural and related to perceptions of quality of care)
4. Improvements in quality – particularly in relation to newborn care and also in terms of improving the experience of women who deliver – are crucial, both to increasing use of services and making them effective
5. The right payment systems must be in place (to cover costs but also reward efficiency), but so must effective regulation of facilities and staff, to prevent illegal passing on of costs to users
6. Clear communication, involvement of stakeholders, including providers, and enabling effective stewardship of the local health system are all critical to policy and wider health service effectiveness
This leads to some overall recommendations:
i) Governments should extend exemption and subsidy policies but should ensure that they cover a package of care which addresses all of the main causes of maternal and neonatal morbidity and mortality
ii) They should work with providers to raise the quality of care provided, particularly for the newborn, and not neglecting the interpersonal skills which are so important to users
iii) Policies need to be clearly communicated to health staff and the community
iv) Provider payments need to be correctly calibrated so as to create the right incentives
v) Managers, staff and communities should be involved in developing and monitoring the policy in order to increase ownership and control abuse
vi) All countries need to more effectively regulate providers to stop illicit payments from being demanded of patients
vii) Additional actions are needed to ensure that benefits can be equitable – in particular, paying attention to transport and access to facilities at night
viii) Underlying systemic weaknesses which undermine policy effectiveness need to be addressed. These include, for example, drugs supply and distribution systems which are not reliable and poor provider-patient relationships.
ix) There should be a focus on enabling effective stewardship at the local level by reinforcing competences and setting up institutional arrangements to enable positive management of resources.
x) All exemption and financial protection policies should be embedded in an overall national plan to achieve universal health coverage, and should not add to the fragmentation of the health financing architecture.
FEMHealth has produced a series of reports to date, most of which are available on the project website: www.abdn.ac.uk/femhealth. These will continue to be disseminated after the end of the project.
WP2 HEALTH FINANCING REPORTS
Financial flows tracking reports
Agbofoun, T. and Makoutode, P. Les résultats de l’enquête sur le suivi des flux financiers, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'enquête sur le suivi des flux financiers, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. Les résultats de l’enquête sur le suivi des flux financiers, Maroc, FEMHealth.
Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur les flux financiers, Mali.
Costing reports:
Makoutode, P.C. Lawin, E., Agbla, S.C. Les résultats de l’étude de coût, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'étude de coût, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Les résultats de l’étude de coût, Mali, FEMHealth.
Exit interview reports:
Makoutode, P.C. Lawin, E. Agbla, S.C. Les résultats des entretiens à la sortie, Bénin, FEMHealth
Ganaba R., Ilboudo P., Cunden N., Witter S. Résultats de l'enquête de l'analyse du coût supporté par le ménage pour l'accès aux soins de l'accouchement et des soins obstétricaux d'urgence au Burkina Faso, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. L’impact de la politique sur les dépenses des ménages, Résultats des entretiens à la sortie, Maroc, FEMHealth
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur l’analyse des dépenses ménages, Mali.
Health worker survey reports:
Makoutode, P.C. Lawin, E. Agbla, S.C. Les résultats de l’enquête des personnels de santé, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'enquête sur le personnel de santé, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. Les résultats de l’enquête des personnels de santé, Maroc, FEMHealth.
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur la motivation du personnel de santé, Mali.
WP2 Health Policy
Filali, H. La politique de gratuité des accouchements et césariennes au Maroc
Filali, H. Rapport de l'analyse sur la politique de gratuité des accouchements et césariennes au Maroc
Yaogo, M. Synthèse des résultats du PA2 au Burkina Faso
Affo, F. Politique de la gratuité de la césarienne au Bénin: faits, écueils et attentes des acteurs: évaluation socio-anthropologie d'une action publique
Lange, I. And Witter, S. (2013) Put them together, shake it, get New perspectives: Case study of the Financial Access to Health Services Community of Practice and its workshops.
WP3
Van der Veken, K., Richard, F., Marchal, B., Witter, S., Dossou, J-P., Essolbi, A., Yaogo, M., Dubourg, D. & De Brouwere, V. POEM – Policy Effect Mapping (2014). A framework to assess the effects of a targeted policy on the local health system.
Marchal, B., Van der Veken, K., Essolbi,A., Dossou, J.P. Richard, F. and Van Belle,S (2014). Methodological reflections on using realist evaluation in a study of fee exemption policies in West Africa and Morocco
Marchal,B., Van Belle, S., De Brouwere,V., Witter, S. & Kegels, G. Complexity in health (2014). Consequences for research & evaluation. FEMHealth discussion paper.
WP3 national POEM reports
Yaogo M., Diallo C., Diallo B., Richard F., Marchal B., De Brouwere V., 2014. De l'implantation à la mise en oeuvre de la subvention nationale des accouchements et des SONU : des effets contrastés dans l'application dans 6 districts évalués, Rapport d'analyse POEM, FEMHealth, Bobo-Dioulasso.
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Richard, F., Marchal, B., De Brouwere, V. 2013. Evaluation du degré de mise en oeuvre de la politique de gratuité de l'accouchement et de la césarienne au Maroc. FEMHealth, Rabat.
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Richard, F., Marchal, B., De Brouwere, V. 2014. Cartographie des effets de la politique de gratuité de l'accouchement et de la césarienne au Maroc. FEMHealth, Rabat.
Dossou JP, Tonouheoua O, Kanhonou L, Goufodji S, Makoutodé P, Agbla S, Lawin L, Richard F, Marchal B, Dubourg D et de Brouwere V. Cartographie des Effets de la Politique de Gratuité de la Césarienne sur les Systèmes Locaux de Santé au Bénin : Rapport de recherche du projet FEMHealth, Bénin, 2014
WP3 national RE reports
Dossou JP, Marchal B, Tonouheoua O, Kanhonou L, Goufodji S, Makoutodé P, Richard F, de Brouwere V. Mise en œuvre de la politique de gratuité de la césarienne au Bénin : A-t-elle réussi ? Pour qui, dans quel contexte et comment ? Rapport des études de cas réaliste, Projet FEMHealth, Bénin, 2014
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Marchal, B., De Brouwere, V. 2013. Evaluation Réaliste de la politique de gratuité de l'accouchement et de la césarienne : Etudes de cas au Maroc.
Yaogo M., Richard F., Diallo B., Marchal B., De Brouwere V., 2014. Mise en oeuvre de la subvention nationale des accouchements et des SONU : une politique diversement adaptée au système de santé local (cas de 2 districts évalués), Rapport d'analyse sur l'évaluation réaliste, FEMHealth, Bobo-Dioulasso.
WP4
Lange, I. and Kanhonou, L. (2014) The costs of 'free': patients and health workers experiences with c-sections during the c-section fee exemption policy in Benin. Draft article.
Cresswell J, Assarag B, Meski FZ, Filippi V, Ronsmans C (2014) Health facility deliveries and caesarean sections by wealth quintile in Morocco between 1987 and 2011. Draft article.
Kanhonou, L. (2013) Qualité des soins dans le contexte de gratuite de la césarienne au Benin: volet anthropologique du PA4
Ababou, M. (2013) Rapport d'analyse des observations et des interviews sur la qualité de prise en charge des parturientes a Sidi Kacem et El Haouz
Filali, H. (2013) WP4 Maroc analyse qualitative
WP5
Politiques d’exemption pour les services de santé maternelle en Afrique: évaluation, expériences et partage des connaissances: Rapport d’Atelier de Ouagadougou, Février 2014
Avancer avec l’apprentissage » : Rapport du 1er atelier d’évaluation technique et de vision stratégique des facilitateurs et promoteurs des Communautés de Pratique de Harmonization for Health in Africa, Allison Kelley et Isidore Sieleunou, Novembre 2013
Equity in Universal Health Coverage: How to Reach the Poorest: Marrakesh Workshop Report, November 2012
Amélioration de l’accès financier aux soins de santé : quels peuvent êtres les apports du financement basé sur la performance ? Rapport d’atelier de Bujumbura, Juin 2012
Assessing Communities of Practice in Health Policy: a conceptual framework, Bertone et al., March 2012
Technical workshop on the benefits package for maternal health fee exemptions: Bamako Workshop Report, March 2012
Maternal health fee exemptions: Policy Brief, Yamba Kafando, Isidore Sieleunou, Fabienne Richard, Allison Kelley, Sophie Witter, and Guy Clarysse, July 2012
Kelley, A. and Meessen, B. (2014) The Financial Access to Health Services Community of Practice: lessons learned after three years.
COUNTRY REPORTS
Boukhalfa, C. et al.(2014) Evaluation of the impact of the free care for deliveries and caesareans in Morocco. FEMHealth country report.
Goufodji, S. et al. (2014) Synthesis report: evaluation of the free caesarean policy in five health districts, Benin. FEMHealth country report.
Ganaba, R. et al. (2014) Cost and impact of the subsidy policy for deliveries and emergency obstetric care in Burkina Faso. FEMHealth country report.
Four country protocols are also available on the website.
POLICY BRIEFS
FEMHealth (2014) Why and how fee exemption policies are adopted by district-level health managers: Methodological lessons from a series of realist case studies in Benin, Burkina Faso and Morocco.
FEMHealth (2013) POEM (Policy Effects Mapping tool): how to evaluate the effects of a targeted policy on the local health system.
FEMHealth (2014) Health care near-miss – indicators to measure the performance of obstetric teams in poor resource settings.
FEMHealth (2014) Managing knowledge for better health policies: the Financial Access to Health Services Community of Practice experience.
Kafando,Y., Richard,F., Kelley,A., Sieleunou,I., Witter, S. and Clarysse, G. (2011) 'Maternal health fee exemptions- Policy Brief' Workshop on the benefits package for maternal health fee exemptions November 17-19, 2011 Bamako.
FEMHealth (2014) Evaluation of the impact of the free care for deliveries and caesareans in Morocco.
FEMHealth (2014) Synthesis report: evaluation of the free caesarean policy in five health districts, Benin.
FEMHealth (2014) Cost and impact of the subsidy policy for deliveries and emergency obstetric care in Burkina Faso.
FEMHealth (2014) Free care for deliveries and caesareans in West Africa and Morocco: research findings and lessons.
CONFERENCE PRESENTATIONS
FEMHealth helped organise an initial conference to debate issues around the design of maternal fee exemption packages. This was held under the aegis of the Community of Practice in Bamako in November 2013: http://www.abdn.ac.uk/femhealth/current/community-of-practice/
FEMHealth made a series of methodological conference presentations at the Health Systems Research conference in Beijing in 2012:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/beijing-2012-global-symposium/
Initial findings from the country studies and on the methodological innovations were presented at the closing workshop, co-hosted with two universities, in Ouagadougou in November 2013:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/ouagadougou-2013/
Each of the main country findings were also presented at the African Health Economics Association meeting in Nairobi in March 2014: http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/afhea/
In-country dissemination events were held early in 2014 – see, for example, from Morocco:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/morocco/
OTHER DISSEMINATION ACTIVITIES
Stakeholders at the country level were identified as follows:
• National level policymakers and advisors (Ministry of Health, Ministry of Finance etc)
• Professional societies of obstetricians/gynaecologists, midwives etc.
• Managers who implement policy at national and district levels
• Health care workers who are responsible for providing services
• Non-governmental organisations (NGOs) and civil society organisations (CSOs) involved in maternal health and other intermediaries, such as the media
• Research participants and their wider communities
Findings have been disseminated to them through a series of feedback sessions in-country, through bilateral meetings with key partners at national level, through the distribution of summary briefs, through steering committee meetings, as well as inviting key technical staff to participate in FEMHealth and Community of Practice conferences.
Stakeholders at regional and international level were identified as the following:
• Policymakers and advisors working in multilateral organisations
• Donor organisations
• Researchers and academics
• International NGOs
FEMHealth has disseminated findings to them through a variety of channels, including publication of articles in international journals; emailing of research summaries to a wide range of networks including the Community of Practice members and maternal health networks; writing blogs to publicise key points arising from the research (e.g. for the id21 site, HEART site, Guardian newspaper and others, ongoing), and circulation of findings within academic networks. International organisations have been engaged through CoP meetings, through involvement in the FEMHealth advisory board, and through other channels, such as virtual networks. The CoP website and FEMHealth website have been key repositories for reports, news and findings.
List of Websites:
http://www.abdn.ac.uk/femhealth/
Mr. Alec Cumming
UNIVERSITY OF ABERDEEN
Tel:+441224438107
Fax:+441224438110
E-mail:alec.cumming@abdn.ac.uk
This report summarises some of the main conclusions and lessons from multi-disciplinary evaluations of four national policies to increase access to obstetric care in Burkina Faso, Benin, Mali and Morocco. These countries have all introduced universal subsidy or exemption policies targeting deliveries and caesareans (in some cases) in the last decade. The FEMHealth project used a variety of research tools to understand the origins, financing, implementation and effects of these policies, focussing not only on utilisation, health gains and financial protection but also on impacts on the wider health system and any unintended effects.
There is a growing consensus that maternal health outcomes can only be improved through policies and programmes that combine interventions to address the different causes of ill health and target multiple groups. Such policies and programmes are complex in nature as they involve coordination between different tiers in the health system and multiple actors including communities, health workers and managers. User fee exemption for delivery and emergency obstetric care (EmONC) is one such policy that has been introduced by many countries in Africa and globally with the aim of improving access to care and thus improving maternal and neonatal outcomes. However, the current evidence base regarding the impact of this policy is not well developed, in part because of evaluation designs that are not able to capture all the necessary information for policy-makers to make informed decisions. The FEMHealth project (2011-14) aimed to reduce this gap by developing research methodologies and tools that would lead to enhanced research on policy implementation, stronger evidence and improved dissemination.
The findings presented in this report are taken from 14 main research tools, most of which were used in all four study countries. They used mixed methods and included: document review; interviews with key informants at international, national and district level; analysis of routine financial and health information, as well as of secondary survey data; structured extraction from medical files; surveys of patients and staff; in-depth interviews with patients and observation of care processes. Conceptual frameworks linking all of these components were developed at the start of the programme and helped to integrate results.
Within each country, 4-6 study sites were chosen using criteria relating to minimum samples, but also a range of contextual criteria including poverty rates, utilisation levels, and distribution of population - the aim being to include a variety of contexts, including areas with higher barriers of poverty and low utilisation as well as ones where access was better prior to the policy introduction.
In order to study effects, FEMHealth used previous studies, where available, and time trend analysis. We also used cross-sectional analysis to understand how differences in outcomes might link to differences in implementation of the policies. Our overall approach was to construct a plausible narrative of policy introduction, implementation and effects, based on a triangulation of different sources and methods, and to observe any patterns across the different sites and contexts.
Key findings were as follows:
1. All three countries (recent data was lacking for Mali) have seen increasing access over time, and while the policies may have contributed to their continuation into the current period, they have not apparently accelerated that trend
2. Analysis of changing utilisation by socio-economic group shows a narrowing of inequalities for all three countries with survey data. Again, this is part of a longer term trend and relates in part to the prior high supervised delivery rates of higher income quintiles. It is not possible to quantify the role of the policies, though they are likely to have contributed to some extent
3. The overall evidence suggests a significant reduction in household payments for the targeted services, ranging from 60 to 90%, depending on the delivery type and country
4. However, even in relation to the package of care which was supposed to be covered, households continued to pay sums which amounted to a varying proportion of their overall expenditure (small in Morocco, intermediate in Burkina and Benin and substantial in Mali)
5. Cross-sectional analysis shows that quality of care is variable across sites and that there is no evidence that those hospitals which are implementing the policy effectively are providing worse care. In some cases, such as the Burkina Faso sites, it is the opposite pattern
6. There were no systematic effects on the wider health system or on untargeted services
7. From a financial perspective, the financial burden of the policies is manageable. They cost 2.5-3.5% of public health expenditure in 2011, and were funded from national resources
Some notable strengths were found across the countries. There is considerable political commitment to these policies and to maintaining the gains they have generated. They have been relatively thoroughly implemented, without the budget shortfalls which have been documented in other settings. They have also, in some cases, like Morocco’s, been accompanied by the additional supply-side improvements which are required. Our research also found an underlying support for them, not only from beneficiaries but also from key actors within the health system (health district managers, hospital management teams, specialists, nurses and midwives). The policies were generally considered to be relevant and important. In some cases, they also appeared to generate efficiency savings.
However, the heavy emphasis of these policies on caesareans (in two out of four countries) has been problematic in a number of ways: caesareans can save lives, but even if utilisation increases it is not easy to know if the right (medically indicated) women have received care. Moreover, the use of caesareans is heavily skewed to the rich and to urban areas, meaning that the benefits of the funding will almost automatically be biased in favour of the rich. It is an intervention which in some contexts needs boosting, but in other contexts (or for some groups) needs controlling. It can be induced by suppliers and patients for the wrong reasons, and carries medical risks.
Moreover, in other countries, where fixed payments to facilities were high and where onward charges to women continued illegally under the policy, the main winners were the hospitals, which gained financially from the policy. Unfortunately there was limited evidence of this being used to improve the overall quality of care or to strengthen the health system as a whole.
We have found a range of outcomes in different contexts (positive and negative across different sites within the same country). This underlines the importance not just of policy design but also context and the institutional and organisational frameworks into which policies are introduced.
The policies are relatively recent and it is early to make a final judgement on this question, but in general, they are likely to have played a part in supporting continued improvements, along with other changes and investments. In the case of Morocco, for example, there were a number of parallel investments within the overall Action Plan and many households prior to the policy benefited from a card providing exemption from all payments for low income households. In this context it would not be expected that the free delivery care at hospitals would lead to a dramatic shift in behaviour. Rather, it should be seen as part of a continuum of measures to increase facility deliveries over time.
The policies are universal in design and should benefit all women. However, non-financial barriers are more significant for women in rural areas, particularly in relation to transport. Moreover, the policies support those who use the services, which in all countries were skewed towards the better off households before the policies (especially for caesareans).
Changing that means changing care-seeking behaviour at delivery, which as our research has shown is quite a difficult task, especially over the short term. This requires raising awareness of the policies, especially amongst non-users and more remote women, improved physical access and reassuring women in relation to their reception, the costs they will face and the support they will receive during deliveries.
Evidence suggests that all of these require more effective actions in the study countries. A significant proportion (0-35%, depending on site) was unable to pay, even after policy implementation. Moreover, women reported a lack of certainty about what they should pay or not which not only increases financial problems but also clouds the relationship with providers. This indicates that there is plenty of scope to increase the financial protection offered by the policies. For example, in Burkina Faso, the payment of the residual 20% for indigents remains to be implemented.
The case studies found that the targeted policies have not created opportunities to strengthen the stewardship function, or at least, these opportunities have not been taken. This may present a challenge beyond the scope of a single national policy.
Project Context and Objectives:
PROJECT CONTEXT AND OBJECTIVES
There is a growing consensus that maternal health outcomes can only be improved through policies and programmes that combine interventions to address the different causes of ill health and target multiple groups. Such policies and programmes are complex in nature as they involve coordination between different tiers in the health system and multiple actors including communities, health workers and managers. User fee exemption for delivery and emergency obstetric care (EmONC) is one such policy that has been introduced by several African countries with the aim of improving access to care and thus improving maternal and neonatal outcomes. However, the current evidence base regarding the impact of this policy is not well developed, in part because of evaluation designs that are not able to capture all the necessary information for policy-makers to make informed decisions. The FEMHealth project (2011-14) aimed to reduce this gap by developing research methodologies and tools that would lead to enhanced research on policy implementation, stronger evidence and improved dissemination.
The overall aims of the project were: (1) to develop new methodological approaches for the evaluation of complex interventions in low income countries, (2) to improve the health of mothers and their newborns by performing comprehensive evaluations of the impact, cost and effectiveness of the removal of user fees for delivery care on maternal and neonatal health outcomes and service quality, and (3) to improve the communication of this evidence to policy-makers and other stakeholders. This report summarises evidence in relation to the second objective, drawing across a range of research components and tools.
At the start of the project, a set of core research questions were developed which outlined the main levels, domains and topics which FEMHealth would seek to investigate (Figure 1). These form the broad structure of this report, which flows from an analysis of the drivers behind the policies, their objectives and formulation at the national level to analysis of how they interacted with and impacted on district-level health systems, and, finally, their effects and effectiveness at community and household level.
RESEARCH METHODS
At the start of the project, a set of core research questions were developed which outlined the main levels, domains and topics which FEMHealth would seek to investigate. The first set (within work package 2) relate to health policy and health financing analysis. The second (WP3) focussed on analysis of the health district implementation and effects. The final work package (WP4) analysed changed to utilisation and quality of care. These form the broad structure of this report, which flows from an analysis of the drivers behind the policies, their objectives and formulation at the national level to analysis of how they interacted with and impacted on district-level health systems, and, finally, their effects and effectiveness at community and household level.
The findings presented in this report are taken from 14 main research tools, most of which were used in all four study countries. They used mixed methods and included: document review; interviews with key informants at international, national and district level; analysis of routine financial and health information, as well as of secondary survey data; structured extraction from medical files; surveys of patients and staff; in-depth interviews with patients and observation of care processes. Conceptual frameworks linking all of these components were developed at the start of the programme and have been described elsewhere (Marchal et al. 2013).
Within each country, 4-6 study sites were chosen using criteria relating to minimum samples, but also a range of contextual criteria including poverty rates, utilisation levels, and distribution of population - the aim being to include a variety of contexts, including areas with higher barriers of poverty and low utilisation as well as ones where access was better prior to the policy introduction.
More details on the methods can be found in the constituent reports, which will be made available on the FEMHealth website (www.abdn.ac.uk/femhealth).
All protocols received in-country ethical approval – in October 2011 in Burkina Faso, January 2012 in Morocco, March 2012 in Benin, and July 2012 in Mali. In addition to the approval from the ethics committee, administrative authorization was requested and obtained from the regional health departments, districts, hospitals and at national level in all four countries. The component tools supported by thematic work packages were also approved by the ethics committees at the LSHTM and ITM.
The main limitation faced by the study as a whole was that in each case, policies had been introduced nationwide and some years earlier, with no control areas and no formal baseline data. In order to study effects, FEMHealth used previous studies, where available, and time trend analysis. We also used cross-sectional analysis to understand how differences in outcomes might link to differences in implementation of the policies. Our overall approach was to construct a plausible narrative of policy introduction, implementation and effects, based on a triangulation of different sources and methods.
In the case of Mali, civil conflict has meant that the full data set could not be collected. Consequently, this report contains some results for Mali, but not the full evaluation results which were obtained from the other three countries.
Project Results:
CONTEXT AND DRIVERS BEHIND POLICIES
One of the FEMHealth research strands aimed to understand the origins of the policies in this region – why had so many countries adopted similar policies over a short period? What were the drivers behind this and how far had they been influenced by one another and by international actors?
One set of drivers related to context and the recognition by decision-makers that socio-economic factors were behind low overall supervised delivery rates in Mali, Burkina Faso and Morocco (Benin already had relatively high facility deliveries at 80%) and large inequalities in all four countries. Design of policy varied however: Morocco took the most wide-reaching approach, embedding the free obstetric care at hospital level within a wider action plan which included investments in quality of care and improved governance. Mali and Benin were the most specific, focussing purely on caesarean sections and some related complications. Affordability was one factor behind the narrowing down of services in Benin.
The hypothesis that international actors might have been behind the proliferation of related policies in the West Africa region was not sustained: interviews suggested that while international influences have been important in shaping the global climate which permeate the four case study countries, the decision-making and elaboration of the policies were dominated by local factors. International actors may have lost some credibility through changes in stance on issues like user fees, with strong but changing messages over the past two decades.
Countries which were once the heartland of the Bamako Initiative have been amongst the most active in taking up selective exemptions (which were seen as more acceptable and affordable than broader approaches to fee removal). The emphasis on the MDGs and, now, on universal health coverage have influenced the underlying discussions, and these free care policies have to be seen in a context of proliferating exemptions in the countries and region for many different vulnerable groups and priority services, though variably implemented. Shame at performing less well in relation to neighbours was highlighted as a driving factor for Morocco. Personal political leadership was seen as critical in all contexts to enabling the policies to be realised – particularly for setting out a vision, and mobilising funds and support. Evidence, while it was marshalled quite thoroughly in two countries (Morocco and Burkina Faso), was used more to assist with planning implementation details than in propelling the original policy adoption itself.
In relation to learning across countries, there was some evidence that countries had looked with interest at the experience of their neighbours who had started implementing earlier, but no formal channels of transmission of detailed lessons were identified.
DESIGN AND IMPLEMENTATION OF THE POLICIES
While there are many common elements, the policies differed in some respects. The government in Burkina Faso opted for a subsidy policy supporting 60 to 80% of the costs of obstetric and neonatal care in all public health facilities in the country. This initiative differs from what is in place in Morocco where normal delivery and caesarean sections are totally free at all levels of public hospitals. The policies in Benin and Mali are also different from the Burkinabe one since they are based on 100% reimbursements of caesarean costs alone. In the case of Benin, this covered accredited confessional and private facilities as well as the public sector.
All medical costs associated with the target services within hospitals (and health centres in Burkina) were included in principle in the package. None of the countries covered transport to the first level facility, but all claimed to cover onward referral transport, though in practice patients often paid. For Morocco, food within the hospital was covered, but this was not the case for the other countries.
Implementation arrangements also varied: most policies were managed by a national committee but in the case of Benin, a dedicated autonomous agency was established to manage the policy. These differences are reflected in some of the research findings: for example, the centralised model adopted by Benin may explain why there were no discrepancies or major delays in reimbursement flows to facilities. However, there were also downsides, in terms of a lack of involvement of the zone in managing and monitoring the policy (the chain of command went straight from the agency to the hospitals, without involving district managers).
FINANCING
In all four countries, the policy is financed almost entirely by the state, with a notable absence of direct donor financial support and indirect donor support for the sector ranging from 0.5% of public health spending in Morocco to 36% in Burkina Faso.
The three sub-Saharan African countries reimburse retrospectively according to the number of services provided, using fixed payments per caesarean in Mali and Benin, and reimbursement of actual variable costs in Burkina Faso. For Morocco, hospital budgets have been boosted to reflect rough estimates of revenue likely to be lost due to the withdrawal of fees for deliveries, but these bear no relation to actual workload and have not increased over time (indeed all subsidies for hospitals were unpaid in 2011). Additional investments in staffing and improved access and care were made in Morocco, which was not the case for any of the other countries. These features are reflected in some of the findings on implementation and effectiveness: quality of care scores, for example, are relatively good in Morocco.
The payment systems also create specific incentives. For Burkina Faso, the payment of actual costs means that there is no incentive to over-provide or to skimp on quality, but on the other hand, there is no surplus to reinvest (the costs are just the variable ones, such as drugs and supplies, so there is no contribution to general facility running costs). In addition, the workload implied by billing by item is considerable, which is why this factor caused discontent amongst the staff in Burkina Faso, more than in any other country. By contrast, the fixed tariff in Benin was found to overpay caesareans at all levels of hospital (relative to actual production costs but also relative to previous payments from users). On the positive side, examples of managers reinvesting the surplus in improving overall services were found. On the negative side, there was a perception that staff were sometimes too eager to do a caesarean, rather than letting women try for a normal delivery. In addition, despite the generous payments, some hospitals continued to make women pay for items which should be free (see below).
The scale of investment is hugely different across the countries, with Morocco spending in the region of 24.5 million Euros on its overall action plan in 2011, compared to 3.2 million Euros in Benin, and 415,000 Euros in Burkina Faso. Per delivery covered that equates to 1.3 Euros for all deliveries (Burkina), 152 Euros per caesarean (Benin), and 797 Euros for all deliveries combined (Morocco). Clearly, these differing scales of investment should shape our expectation of results.
As a proportion of public health expenditure in 2011, the policies absorbed around 2.5% in Morocco, 3% in Benin and 3.5% in Burkina Faso – not insignificant but all potentially sustainable, if the policies are seen as effective.
HEALTH SYSTEMS EFFECTS
Human resources
The only country to accompany its policy with a significant increase in staff was Morocco, which increased the deployment of midwives and hospital specialists as part of its overall action plan. For other countries, there was no evidence of increased numbers of staff in any large measure, though in Benin, the additional resources provided by the policy allowed some local changes to improve staffing. Reported working hours for the main groups of staff concerned with obstetric care remain reasonable even after the policy, across all four countries, but workload (patients seen and deliveries attended) are reported to have increased for three out of four of the countries (Mali being the exception). In Burkina Faso, the administrative workload imposed by the policy was a particular concern. The consensus was that the policies have not affected their remuneration (as there are no direct financial incentives linked to it for any of the staff in any of the countries).
Across all countries the majority of staff surveyed feels positive about the wider effects of the policy, believing that the policy has increased access to supervised deliveries, has benefited the poor, and has improved the quality of care, including through improvements to drugs and supplies. This also impacts positively on health worker working conditions and satisfaction in some countries, like Burkina Faso and Mali.
Facility finances
In Morocco, despite the increased budget to accompany the implementation of the free caesarean and delivery policy, the general revenues of the health facilities have not been affected positively. In 2011 and during the first half of 2012, no financial resources were provided at all. This has affected the financial reserves of the health facilities that are still implementing the gratuity policy. For Benin, the costing revealed that the current tariff is beneficial to hospitals, paying considerably more than it costs to provide the service, though the surplus varies by type of hospital. In addition, payments have been regular. For Mali, the calculus is different, with the reimbursements not fully covering costs of provision. In Burkina Faso, payments are often delayed and do not include any operating costs beyond the direct inputs needed for the service.
It might be expected that extra costs would be levied from users in countries where facilities are poorly rewarded for providing care, however analysis of household payments does not support this. In absolute terms, women reported paying the largest amount overall for caesareans in Benin (60 Euros on average), where the payment for the policy is most generous. This suggests that organisational culture and other factors play a role in the levying of additional payments from users, rather than actual financial need from the facility perspective.
IT systems
There has been no broad effect of the policies on health information systems, either positive or negative. In Benin, there has been a policy-related improvement in the obstetric information gathering, but this has not cascaded into other areas of health information. In Burkina Faso, a specific system of collecting information on the policy was put into place, and it is operating with variable success. As for Morocco, the wider information system is unchanged and is generally seen as onerous for health staff.
Drugs and supplies
All four countries used kits to support the implementation of the policy, with varying results. The policy in Morocco was accompanied by a large increase in kits, which meant that drug supply improved, though in some places, kit numbers were well in excess of need. For Burkina Faso, drug supply also improved, but management of kits sometimes lacked transparency and stock-outs did sometimes occur. In both Morocco and Burkina Faso, equipment was felt to be lacking, with no provision made to improve working conditions. In Benin, the kit contents were perceived to be too generous at first, leading to misuse, and then too restrictive, leading to costs to users for items not included in the kits. In general, it could be said that attention was paid to providing specific drugs and items needed for obstetric care, but there was no wider investment in the supply system and a continuing focus on the use of kits, which as seen, may not be the most efficient method of organising supplies.
Management
The policies introduced threats and opportunities for local managers, and examples were found of positive and negative adaptation. Generally, managers had not been much involved in the development of the policy, and in many cases detailed guidance on how to implement it was lacking. Management teams have leeway to interpret the policy. In some cases, it was found that actors such as directors and specialists used their power position to adapt the policy to their own benefit, e.g. in some study sites of Benin or Morocco, patients were still charged fees, because staff were compensating for the free caesareans by charging something else to the patient. These effects can be moderated or avoided by capable management teams and adequate supervision by programme managers, as seen in other study sites of the same countries, where district health managers and or hospital directors are positively engaged with the policy implementation and the protection of service users.
In most study sites in Benin, Mali and Burkina Faso, the management teams did not have a large absorption capacity to take up the new tasks without additional resources. When the reimbursements were late or inadequate, the implementation halted. In Morocco, this was not the case, as the capacity of the teams, the existing implementation infrastructure in terms of human resources, facilities, and equipment was adequate to take up the new patients.
FACTORS BEHIND DIFFERENTIAL IMPLEMENTATION
The analysis of divergent responses of different cadres within our realist evaluation case studies indicated that the adoption of the policy is explained in part by the configuration of autonomy, decision space and motivation of these actors and by organisational, institutional factors and contextual factors. Different responses were found by teams in different sites within the same country, indicating that beyond policy design and national features, there is an inter-play of local factors which influences whether the policy is blocked, adopted or adapted. This highlights the importance of reinforcing the stewardship function, not only in relation to these policies but more generally. The role of ensuring public accountability, in particular, was found to be underdeveloped or even not mandated clearly. In any case, lack of effective stewardship allowed faulty implementation processes to continue in many of the sites.
IMPACT ON UTILISATION
• For Burkina Faso, there is a statistically significant decline in the rate of increase facility deliveries (i.e. between 2002 and 2007 there is a 12% relative increase each year, whereas between 2007 and 2010 there was a 6% relative annual increase) - so whilst the overall rate is increasing in absolute terms, it is doing so less rapidly post-2007. For caesareans, there is no evidence of a change in either direction.
• For Morocco, there is no evidence of a change post-2008 in facilities deliveries: the gradient is flat. There is actually a significant decrease in caesarean rates after 2008, though this may be related to the 2008 data point being erroneously high. In either case, there is no evidence of a change in gradient.
• For Benin, there is an upward trend but no significant change in gradient post-policy
• For Mali, we have no comparable post-policy data, though the change may be modelled when the next DHS dataset becomes available.
It is evident overall that countries have made progress over the past 15-20 years, and these policies may have contributed, but there is no evidence of that as yet. It is almost certainly too early to tell, as we have only 2-3 post-policy data points in each country, and the varying implementation documented by FEMHealth also underlines the need to be cautious about assuming immediate effectiveness of policies.
Furthermore, the Morocco caesarean section rate is at a level where any increase (even if true) would be unlikely to lead to further reductions in maternal mortality, and if anything the concern is excess and unnecessary caesareans among certain groups.
IMPACT ON OTHER (UNTARGETED) SERVICES
The research looked in a systematic way for the impact of the policies on untargeted services, in order to capture any unintended positive or negative effects, but found no major effects. Some positive effects of additional resources introduced by the policies were documented (e.g. in Morocco and Benin), as well as from wider utilisation uptake in Burkina Faso, linked to the policy. On the negative side, some examples were found where the policy encouraged resources (such as staff) to move from untargeted to targeted services. More significantly, in Benin, there was some evidence of supply-induced demand for caesareans, at the expense of normal deliveries. Trend analysis of provision of general medicine and paediatric services in all countries and sites did not reveal any evidence of distortions linked to the exemption policies.
IMPACT ON QUALITY OF CARE
Overall, women’s perceptions of the quality of the services were generally high or very high, and did not correlate well with technical quality of care scores (being highest in Burkina Faso, where technical scores were lowest). The quality of neonatal care, measured by the number of omissions in routine neonatal procedures, was very poor in some hospitals in Benin and Burkina Faso, and generally poorer than the quality of maternity care. Median delays in receiving caesarean sections were above the expected threshold of 1 hour in most hospitals (except in Morocco) and were the highest in Benin hospitals where the policy was designed to facilitate access to life saving emergency surgery. Hospitals in Morocco performed consistently better than hospitals in the other countries.
In observations and interviews in Benin and Morocco, incidents of poor communication were observed, including lack of informed consent for surgical care and poor bedside manners. In Benin, while poor interpersonal relationships in maternity wards had been documented in previous studies, in this project these were attributed by several respondents to perverse effects of the free caesarean policy.
Hospitals are still receiving many cases of near-miss, particularly maternal near-miss, with hospital incidence ranging from less than 2% in Morocco to 14% in Benin. As cases of maternal near-miss are women who nearly died and were saved in extremis, there is still a lot of progress to be made in the organization of the health services in order to reduce the burden of several morbidity and mortality in the focus countries.
Our key hypotheses included that hospitals/districts with lower user fees cost may register shorter delays and fewer adverse events because women may arrive earlier in facilities; but that on the other hand, an increase in volume of patients, if not met with an increase in human resources, might lead to a deterioration of the quality of maternity and neonatal care. While in Burkina Faso, there seemed to be a positive correlation between average omission scores, average delays and the success of the implementation of the policy (as measured by excess payments made by patients under the policy), limited relationships exist between the omission and implementation score in Benin, implying that quality of the care provided was affected by many factors which may be quite independent from policies designed to increased access.
IMPACT OF POLICY
Awareness of the policies was relatively low, ranging from 20% amongst women who had delivered in Benin to 53% in Mali (but much lower in some districts). As these interviews were conducted with women who had used services and had already delivered, wider awareness can be assumed to be even lower. Detailed knowledge of entitlements was very low. Clearly, if policies are to influence care-seeking a greater communication effort is needed, especially for poorer and more remote women (awareness tended to rise for higher quintiles).
Impact on delay in seeking care and health seeking behaviour
In general, median reported delays in leaving home, arriving at facilities and being seen were acceptable, although there was a large variation between sites and by type of delivery, and a wide range within the responses in general. The median delay in leaving the house was rather high in Benin. In addition, it is important to remember that those interviewed represent those who were able to access care, and these are often not the most disadvantaged women. Perceptions of need, the availability of transport and the availability of the key decision-maker (who varied by site) emerge as significant.
In-depth interviews with users suggest an appreciation of the policy and that the policy did address some of the key barriers to access. However, it did not necessarily change health seeking behaviour. In Morocco, interviews showed that the choice of location for deliveries was made largely according to expected comfort, care and monitoring (for example, at home), and reassessed in cases where outside help was decided to be necessary as a matter of urgency. Transport was a key barrier. At the hospital, the absence of a doctor, the gaps in surveillance, inadequate resources, and tedious negotiation process to receive the desired care were all aspects known and anticipated by women when considering their recourse to care.
In Benin, there was also no evidence from the interviews or observations that women modified their decision to seek out skilled professionals at birth due to the policy. While the policy was appreciated and the costs of caesarean sections were considered to be more affordable than in the past, the policy does not erase the fear of the caesarean section as a medical procedure and the threat of loss of life. Decisions about where to give birth were based on hospital reputation, convenience of access and past delivery experiences.
Impact on inequities of access
In all three countries for which there is recent household survey data (i.e. excluding Mali), the relative inequity between the poorest and the richest has declined over time (in that there have been bigger gains among the poorest). The policy may have contributed to this but this is a longer term trend, and one which followed to some extent from the fact that richer women already had high coverage. In all three countries there remains substantial inequity in utilisation of care.
Financial impact for households
In relation to the amounts which should have been paid by households under the policies, we found that they are paying excessive amounts in all countries, though the excess payments are relatively low for Morocco (2% of their total payment for caesareans and 6% for normal deliveries), which indicates a relatively effective implementation of the policy. By contrast, in Mali, 49% of the household payment is excessive. Intermediate proportions of 13% (for Benin) and 17% (for Burkina) were found for caesarean sections. The absolute amounts paid in Benin were higher than for Burkina, as the overall payments were higher.
Patterns of payment across quintiles and across rural and urban areas varied. Large proportions of the households in all settings and quintiles incurred catastrophic costs, even under the current policies (and a higher proportion in rural than urban areas in all four countries).
Certain costs, including transport, cost of companions, care of newborn (in some countries), tipping of health workers and supplementary drugs remain a burden. Lack of clarity on charging can also reduce predictability of costs and cause anger and confusion for women and their families.
Between 0 and 35% (depending on the site) of households surveyed had been unable to make the requisite payments, with those in lower quintiles more likely to report this than higher. In general households coped by using savings and getting help from family and friends, but some had to sell productive assets such as land to cover the bills. Very few had health insurance membership and in general, health insurance was not a protective mechanism for most households, as it is pro-rich in its distribution.
Looking at the difference between recorded payments prior to the policies and average payments now, households have made a substantial financial gain. In Burkina Faso, there was a reduction of 71% for deliveries of all kinds. In Morocco, the gain was lower for normal deliveries (62%), compared to 92% for caesareans. The estimated saving for caesareans in Benin was in the region of 74%, compared to 78% for Mali.
CONCLUSIONS
1. Did the policies increase access to obstetric care?
Analysis of secondary survey data, looking at trends over a longer time span, did not demonstrate a significant increase in utilisation associated with the start of the exemption or subsidy policies in any of the three countries for which we have data (with the exception of Mali). All three countries have seen increasing access over time, and while the policies may have contributed to their continuation into the current period, they have not apparently accelerated that trend.
The policies are relatively recent and it is early to make a final judgement on this question, but in general, they are likely to have played a part in supporting continued improvements, along with other changes and investments. In the case of Morocco, for example, there were a number of parallel investments within the overall Action Plan and in addition, many households prior to the policy benefited from a card providing exemption from all payments for low income households . An estimated 70% of the population were covered by these cards. Although the cards were not fully respected or used (Hodgkin, Krasovec K., El-Idrissi, Eckert, & Karim 2005), in this context it would not be expected that the free delivery care at hospitals would lead to a dramatic shift in behaviour. Rather, it should be seen as part of a continuum of measures to increase facility deliveries over time.
2. Which groups benefited most?
Analysis of changing utilisation by socio-economic group shows a narrowing of inequalities for all three countries with survey data. Again, this is part of a longer term trend and relates in part to the prior high supervised delivery rates of higher quintiles. Gains, where they occurred, were likely to be amongst those with lower use at the start. It is not possible to quantify the role of the policies, though they are likely to have contributed to some extent.
The policies are universal in design and should benefit all women. However, non-financial barriers are more significant for women in rural areas, particularly in relation to transport. Moreover, the policies support those who use the services, which in all countries were skewed towards the better off households before the policies came into effect, and even more so for caesareans. This inevitably means that a disproportionate part of the benefits are captured by better off households. A simple breakdown of our exit interview sample by quintile, across the countries, demonstrates this (Table 23). Relatively few of the users were in the poorest quintile. For policies focussed on caesareans, the three better of quintiles hold the bulk of users. For those with broader policies, there is more evidence of benefits for the poor (e.g. quintile 2).
Changing that means changing care-seeking behaviour at delivery, which as our research has shown is quite a difficult task, especially over the short term – it means raising awareness of the policies, especially amongst non-users and more remote women, ensuring physical access and also reassuring them in relation to their reception, the costs they will face and the support they will receive during deliveries. Evidence suggests that all of these require more effective actions in the study countries. In Burkina Faso, the payment of the residual 20% for indigents remains to be implemented.
3. Were they effective in reducing financial burdens?
The overall evidence suggests a significant reduction in household payments for the targeted services, ranging from 53% for caesareans in Benin to 92% for caesareans in Morocco. Some of the costs were expected to remain as the free care policies focus on in-facility costs. However, even in relation to the package of care which was supposed to be covered, households continued to pay sums which amounted to a small proportion of their overall expenditure in Morocco, intermediate in Burkina and Benin and the large majority in Mali. A significant proportion (0-35%, depending on site) was unable to pay. Moreover, women reported a lack of certainty about what they should pay or not which not only increases financial problems but also clouds the relationship with providers. This indicates that there is plenty of scope to increase the financial protection offered by the policies.
4. What is the overall evidence on effects on quality of care?
Quality of care is complex to measure and baseline data was not available to measure trends over the period (and even with baseline data, attribution to the policy would not have been possible, as quality is affected by so many factors). However, cross-sectional analysis has allowed us to conclude that the quality is variable across sites and that there is no evidence that those hospitals which are implementing the policy effectively are providing worse care. In some cases, such as the Burkina Faso sites, it is the opposite pattern. We can therefore conclude that the policy has not systematically weakened quality. Qualitative evidence has uncovered a large number of positive and negative effects. The realist evaluation case studies suggest that whatever the policy designs, there is scope at the local level to adopt the policies in a positive way. The focus should be on reinforcing competences and institutional arrangements to enable such positive management of resources.
5. Did the policy strengthen or weaken the local health system?
The POEM studies generally indicate that the policy did not reinforce any of the non-targeted services. Often, it led to improvement of resource availability for targeted services, changes in the health information system or competences of providers, but not beyond the sphere of maternal and child health, and specifically maternity and theatre. No changes were documented in general management or organisation of services. POEM and realist case studies of policy adoption have shown that there is no effective stewardship function in the majority of the sites, defined as coordination, management and regulation of all health actors and safeguarding the public interest. The targeted policies have not created opportunities to strengthen the stewardship function, or at least, these opportunities have not been taken. This may present a challenge beyond the scope of a single national policy.
6. What were the main strengths and weaknesses of the policy?
A detailed analysis of strengths and weaknesses are presented in the country reports. However, as a group, certain features stand out. On the positive side, and compared with the documented experience of other countries with similar exemption policies:
• The policies have been relatively thoroughly implemented: despite some gaps and lapses, the policies have been put into effect in a serious way
• They have not been affected by budget shortfalls, which undermined the effectiveness of similar policies in countries like Ghana (Witter & Adjei 2007)
• They have, in some cases, like Morocco’s, been accompanied by the additional supply-side improvements which are required to meet the additional demands
• There is an underlying support for them, and not only from beneficiaries: most actors within the health system (health district managers, hospital management teams, specialists, nurses and midwives) reacted positively to the policy in interviews. The policy was generally considered to be relevant and important
• They have achieved substantial reductions in household payments, which will over time contribute to poverty reduction and reduced inequalities of access
However, there are also some weaknesses. These include:
• A package of care which in some cases (Benin and Mali) will not address all of the main causes of maternal and neonatal morbidity and mortality, and whose impact on these can therefore only be expected to be modest, even if well implemented
• Poorly calibrated provider payments, for those using fixed payments, which either over-incentivise (in the case of Benin) or under-fund (in the case of Mali). Both of these result in perverse effects
• Lack of clear and well disseminated operating documents, which enable staff and clients to be clear about how the policy will work and what is covered by it
• Too limited attention to the quality of care offered by the facilities covered by the policy; for newborns in particular it has been found to be sub-standard
• Lack of involvement in most cases of managers, staff and communities in developing and monitoring the policy in order to increase ownership and control abuse
• No policy has completely reduced the officially exempted costs to zero; although magnitudes of unwarranted payments vary in scale, all countries need to more effectively regulate providers and stop illicit payments from patients
• By their very design, the policies are unable to address some of the main barriers faced by women, such as inability to physically access health care; additional actions are needed to ensure that benefits can be equitable
In addition to the weaknesses of the policies themselves, there are underlying systemic weaknesses which undermine policy effectiveness, such as lack of effective stewardship at the local level, drugs supply and distribution systems which are not reliable and poor provider-patient relationships in some areas.
7. Are these policies likely to be sustainable?
From a financial perspective, the financial burden of the policies is manageable. They cost 2.5-3.5% of public health expenditure in 2011, and were funded from national resources. There is considerable political commitment to maintaining the gains they have generated. The main challenge for sustainability will be developing national plans to move toward universal health coverage which can integrate the patch-work of different exemption and financial protection policies which are now multiplying in all of the study countries. In the course of this, the lessons learned through the FEMHealth evaluation should be borne in mind in order to maximise benefits and minimise risks.
8. Are they cost-effective overall?
The policies in Burkina Faso, Benin, Mali and Morocco were strong national initiatives which aimed to improve maternal health and to increase access to obstetric care. On one level, the evaluation produces inconclusive results: we observe positive trends in relation to supervised deliveries and caesarean sections and a narrowing of inequalities in all three countries for which recent data is available but cannot attribute these to the policies. There is no significant change in the trends which coincides with the introduction of the policies. It is likely that they have contributed to the ongoing trend, but this is only speculative.
The heavy emphasis of these policies on caesareans (in two out of four countries) has been problematic in a number of ways: caesareans can save lives, but even if utilisation increases it is not easy to know if the right (medically indicated) women have received care. Moreover, the use of caesareans is heavily skewed to the rich and to urban areas, meaning that the benefits of the funding will almost automatically be biased in favour of the rich. It is an intervention which in some contexts needs boosting, but in other contexts (or for some groups) needs controlling. It can be induced by suppliers and patients for the wrong reasons, and carries medical risks. While policies to reduce the costs of caesareans (which are high cost, potentially catastrophic events from a households perspective) can provide real financial benefits to households, as these have done, from a public health perspective a wider policy covering a range of life-threatening obstetric complications and also the pathway to them (facility deliveries) is preferable. Morocco and Burkina Faso illustrate this approach.
On that basis, should we conclude that the policies were not cost-effective, or did not represent value for money for the national budget? This requires a more nuanced judgement. One of the underlying objectives was to reduce the burden on households of this essential service, and so change behaviour. The evidence of the FEMHealth research suggests that while financial barriers are significant and are connected to many other barriers (physical, cultural etc.), on their own their reduction does not change behaviour, unless it is connected with a positive shift in other aspects, such as perception of quality and responsiveness. Policies on financial access therefore need to be designed with improvements to these other facets in mind, as an integral part of their design. Behaviour change also has to be measured over a longer period, as habits in relation to significant services such as delivery change slowly.
Looking at a simple comparison of the funds spent by government on the policies versus the estimated gains made by households gives another insight into the value for money question. In Burkina Faso, the average expenditure per delivery is lower than the average gain per household with a delivery. There is therefore a net gain, which probably reflects the payment system and the fact that facilities are providing care without fully recovering their costs. If they are able to do this and still provide adequate care without passing additional costs to women, then the policy is leveraging an efficiency gain in the health system. This is indicated by our results. For Mali, the net balance for caesareans is also positive, though some gaps in our data do not allow us to interpret fully how this has impacted on the health system. For Morocco, the unit cost information from the government side is missing.
For Benin, the calculus is different as the government is paying 100,000 FCFA per caesarean and our results suggest a household gain in terms of reduced costs of around the same magnitude. Societal gains might have been made if the policy had not set a fixed tariff which was too high for average production costs and then also failed to control additional payments which were levied by some hospitals on women and their families. Some of the winners in this scenario are the hospitals which have gained financially from the policy, but unfortunately there is limited evidence of this being used to improve the overall quality of care or strengthen the health system as a whole.
Wider impacts, positive and negative, intended and unintended, are also to be taken into account in coming to an overall judgement about these policies. We have found a range of these but they vary by context and suggest a variety of outcomes can be expected from these policies, depending on their features but also the context and the institutional and organisational frameworks into which they are introduced. This complexity means that no one simple answer to the overall evaluation can be produced for all settings.
The overall recommendations arising from the evaluations are closely linked to tackling the weaknesses outlined above, not only in relation to the policies but also the underlying systemic challenges.
Potential Impact:
Some of the key messages arising from the FEMHealth research are as follows:
1. A broad package of care should be exempted if the goal is improved maternal and neonatal health. Policies which focus on caesareans alone cannot address the main causes of death and disability
2. Exemption and subsidy policies can provide financial protection if well designed and implemented.
3. There are high risks of favouring better off households unless exemption policies are accompanied by concerted efforts to address other barriers (physical, cultural and related to perceptions of quality of care)
4. Improvements in quality – particularly in relation to newborn care and also in terms of improving the experience of women who deliver – are crucial, both to increasing use of services and making them effective
5. The right payment systems must be in place (to cover costs but also reward efficiency), but so must effective regulation of facilities and staff, to prevent illegal passing on of costs to users
6. Clear communication, involvement of stakeholders, including providers, and enabling effective stewardship of the local health system are all critical to policy and wider health service effectiveness
This leads to some overall recommendations:
i) Governments should extend exemption and subsidy policies but should ensure that they cover a package of care which addresses all of the main causes of maternal and neonatal morbidity and mortality
ii) They should work with providers to raise the quality of care provided, particularly for the newborn, and not neglecting the interpersonal skills which are so important to users
iii) Policies need to be clearly communicated to health staff and the community
iv) Provider payments need to be correctly calibrated so as to create the right incentives
v) Managers, staff and communities should be involved in developing and monitoring the policy in order to increase ownership and control abuse
vi) All countries need to more effectively regulate providers to stop illicit payments from being demanded of patients
vii) Additional actions are needed to ensure that benefits can be equitable – in particular, paying attention to transport and access to facilities at night
viii) Underlying systemic weaknesses which undermine policy effectiveness need to be addressed. These include, for example, drugs supply and distribution systems which are not reliable and poor provider-patient relationships.
ix) There should be a focus on enabling effective stewardship at the local level by reinforcing competences and setting up institutional arrangements to enable positive management of resources.
x) All exemption and financial protection policies should be embedded in an overall national plan to achieve universal health coverage, and should not add to the fragmentation of the health financing architecture.
FEMHealth has produced a series of reports to date, most of which are available on the project website: www.abdn.ac.uk/femhealth. These will continue to be disseminated after the end of the project.
WP2 HEALTH FINANCING REPORTS
Financial flows tracking reports
Agbofoun, T. and Makoutode, P. Les résultats de l’enquête sur le suivi des flux financiers, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'enquête sur le suivi des flux financiers, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. Les résultats de l’enquête sur le suivi des flux financiers, Maroc, FEMHealth.
Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur les flux financiers, Mali.
Costing reports:
Makoutode, P.C. Lawin, E., Agbla, S.C. Les résultats de l’étude de coût, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'étude de coût, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Les résultats de l’étude de coût, Mali, FEMHealth.
Exit interview reports:
Makoutode, P.C. Lawin, E. Agbla, S.C. Les résultats des entretiens à la sortie, Bénin, FEMHealth
Ganaba R., Ilboudo P., Cunden N., Witter S. Résultats de l'enquête de l'analyse du coût supporté par le ménage pour l'accès aux soins de l'accouchement et des soins obstétricaux d'urgence au Burkina Faso, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. L’impact de la politique sur les dépenses des ménages, Résultats des entretiens à la sortie, Maroc, FEMHealth
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur l’analyse des dépenses ménages, Mali.
Health worker survey reports:
Makoutode, P.C. Lawin, E. Agbla, S.C. Les résultats de l’enquête des personnels de santé, Bénin, FEMHealth.
Ilboudo P., Ganaba R., Cunden N., Witter S. Résultats de l'enquête sur le personnel de santé, Burkina Faso. 2013. Rapport d'enquête du projet international d'analyse de l'impact des politiques d'exemptions des frais de soins de l'accouchement et des soins obstétricaux d'urgence au Bénin, au Burkina Faso, au Mali et au Maroc ; projet financé par l'Union Européenne.
Boukhalfa, C., Abouchadi, S., Cunden, N. & Witter, S. Les résultats de l’enquête des personnels de santé, Maroc, FEMHealth.
Konaté, K.M. Daou, Z., & Sidibé, K.R. Traore, M.A. Gacko, I. Evaluation de l’impact de la suppression de paiement des frais de soins de santé maternelle en Afrique de l’Ouest et au Maroc : nouveaux outils, nouvelles connaissances. Rapport sur la motivation du personnel de santé, Mali.
WP2 Health Policy
Filali, H. La politique de gratuité des accouchements et césariennes au Maroc
Filali, H. Rapport de l'analyse sur la politique de gratuité des accouchements et césariennes au Maroc
Yaogo, M. Synthèse des résultats du PA2 au Burkina Faso
Affo, F. Politique de la gratuité de la césarienne au Bénin: faits, écueils et attentes des acteurs: évaluation socio-anthropologie d'une action publique
Lange, I. And Witter, S. (2013) Put them together, shake it, get New perspectives: Case study of the Financial Access to Health Services Community of Practice and its workshops.
WP3
Van der Veken, K., Richard, F., Marchal, B., Witter, S., Dossou, J-P., Essolbi, A., Yaogo, M., Dubourg, D. & De Brouwere, V. POEM – Policy Effect Mapping (2014). A framework to assess the effects of a targeted policy on the local health system.
Marchal, B., Van der Veken, K., Essolbi,A., Dossou, J.P. Richard, F. and Van Belle,S (2014). Methodological reflections on using realist evaluation in a study of fee exemption policies in West Africa and Morocco
Marchal,B., Van Belle, S., De Brouwere,V., Witter, S. & Kegels, G. Complexity in health (2014). Consequences for research & evaluation. FEMHealth discussion paper.
WP3 national POEM reports
Yaogo M., Diallo C., Diallo B., Richard F., Marchal B., De Brouwere V., 2014. De l'implantation à la mise en oeuvre de la subvention nationale des accouchements et des SONU : des effets contrastés dans l'application dans 6 districts évalués, Rapport d'analyse POEM, FEMHealth, Bobo-Dioulasso.
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Richard, F., Marchal, B., De Brouwere, V. 2013. Evaluation du degré de mise en oeuvre de la politique de gratuité de l'accouchement et de la césarienne au Maroc. FEMHealth, Rabat.
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Richard, F., Marchal, B., De Brouwere, V. 2014. Cartographie des effets de la politique de gratuité de l'accouchement et de la césarienne au Maroc. FEMHealth, Rabat.
Dossou JP, Tonouheoua O, Kanhonou L, Goufodji S, Makoutodé P, Agbla S, Lawin L, Richard F, Marchal B, Dubourg D et de Brouwere V. Cartographie des Effets de la Politique de Gratuité de la Césarienne sur les Systèmes Locaux de Santé au Bénin : Rapport de recherche du projet FEMHealth, Bénin, 2014
WP3 national RE reports
Dossou JP, Marchal B, Tonouheoua O, Kanhonou L, Goufodji S, Makoutodé P, Richard F, de Brouwere V. Mise en œuvre de la politique de gratuité de la césarienne au Bénin : A-t-elle réussi ? Pour qui, dans quel contexte et comment ? Rapport des études de cas réaliste, Projet FEMHealth, Bénin, 2014
Essolbi, A., Ababou, M., Hachri, H., Van der Veken, K., Marchal, B., De Brouwere, V. 2013. Evaluation Réaliste de la politique de gratuité de l'accouchement et de la césarienne : Etudes de cas au Maroc.
Yaogo M., Richard F., Diallo B., Marchal B., De Brouwere V., 2014. Mise en oeuvre de la subvention nationale des accouchements et des SONU : une politique diversement adaptée au système de santé local (cas de 2 districts évalués), Rapport d'analyse sur l'évaluation réaliste, FEMHealth, Bobo-Dioulasso.
WP4
Lange, I. and Kanhonou, L. (2014) The costs of 'free': patients and health workers experiences with c-sections during the c-section fee exemption policy in Benin. Draft article.
Cresswell J, Assarag B, Meski FZ, Filippi V, Ronsmans C (2014) Health facility deliveries and caesarean sections by wealth quintile in Morocco between 1987 and 2011. Draft article.
Kanhonou, L. (2013) Qualité des soins dans le contexte de gratuite de la césarienne au Benin: volet anthropologique du PA4
Ababou, M. (2013) Rapport d'analyse des observations et des interviews sur la qualité de prise en charge des parturientes a Sidi Kacem et El Haouz
Filali, H. (2013) WP4 Maroc analyse qualitative
WP5
Politiques d’exemption pour les services de santé maternelle en Afrique: évaluation, expériences et partage des connaissances: Rapport d’Atelier de Ouagadougou, Février 2014
Avancer avec l’apprentissage » : Rapport du 1er atelier d’évaluation technique et de vision stratégique des facilitateurs et promoteurs des Communautés de Pratique de Harmonization for Health in Africa, Allison Kelley et Isidore Sieleunou, Novembre 2013
Equity in Universal Health Coverage: How to Reach the Poorest: Marrakesh Workshop Report, November 2012
Amélioration de l’accès financier aux soins de santé : quels peuvent êtres les apports du financement basé sur la performance ? Rapport d’atelier de Bujumbura, Juin 2012
Assessing Communities of Practice in Health Policy: a conceptual framework, Bertone et al., March 2012
Technical workshop on the benefits package for maternal health fee exemptions: Bamako Workshop Report, March 2012
Maternal health fee exemptions: Policy Brief, Yamba Kafando, Isidore Sieleunou, Fabienne Richard, Allison Kelley, Sophie Witter, and Guy Clarysse, July 2012
Kelley, A. and Meessen, B. (2014) The Financial Access to Health Services Community of Practice: lessons learned after three years.
COUNTRY REPORTS
Boukhalfa, C. et al.(2014) Evaluation of the impact of the free care for deliveries and caesareans in Morocco. FEMHealth country report.
Goufodji, S. et al. (2014) Synthesis report: evaluation of the free caesarean policy in five health districts, Benin. FEMHealth country report.
Ganaba, R. et al. (2014) Cost and impact of the subsidy policy for deliveries and emergency obstetric care in Burkina Faso. FEMHealth country report.
Four country protocols are also available on the website.
POLICY BRIEFS
FEMHealth (2014) Why and how fee exemption policies are adopted by district-level health managers: Methodological lessons from a series of realist case studies in Benin, Burkina Faso and Morocco.
FEMHealth (2013) POEM (Policy Effects Mapping tool): how to evaluate the effects of a targeted policy on the local health system.
FEMHealth (2014) Health care near-miss – indicators to measure the performance of obstetric teams in poor resource settings.
FEMHealth (2014) Managing knowledge for better health policies: the Financial Access to Health Services Community of Practice experience.
Kafando,Y., Richard,F., Kelley,A., Sieleunou,I., Witter, S. and Clarysse, G. (2011) 'Maternal health fee exemptions- Policy Brief' Workshop on the benefits package for maternal health fee exemptions November 17-19, 2011 Bamako.
FEMHealth (2014) Evaluation of the impact of the free care for deliveries and caesareans in Morocco.
FEMHealth (2014) Synthesis report: evaluation of the free caesarean policy in five health districts, Benin.
FEMHealth (2014) Cost and impact of the subsidy policy for deliveries and emergency obstetric care in Burkina Faso.
FEMHealth (2014) Free care for deliveries and caesareans in West Africa and Morocco: research findings and lessons.
CONFERENCE PRESENTATIONS
FEMHealth helped organise an initial conference to debate issues around the design of maternal fee exemption packages. This was held under the aegis of the Community of Practice in Bamako in November 2013: http://www.abdn.ac.uk/femhealth/current/community-of-practice/
FEMHealth made a series of methodological conference presentations at the Health Systems Research conference in Beijing in 2012:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/beijing-2012-global-symposium/
Initial findings from the country studies and on the methodological innovations were presented at the closing workshop, co-hosted with two universities, in Ouagadougou in November 2013:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/ouagadougou-2013/
Each of the main country findings were also presented at the African Health Economics Association meeting in Nairobi in March 2014: http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/afhea/
In-country dissemination events were held early in 2014 – see, for example, from Morocco:
http://www.abdn.ac.uk/femhealth/about/impact-and-dissemination/dissemination-events/morocco/
OTHER DISSEMINATION ACTIVITIES
Stakeholders at the country level were identified as follows:
• National level policymakers and advisors (Ministry of Health, Ministry of Finance etc)
• Professional societies of obstetricians/gynaecologists, midwives etc.
• Managers who implement policy at national and district levels
• Health care workers who are responsible for providing services
• Non-governmental organisations (NGOs) and civil society organisations (CSOs) involved in maternal health and other intermediaries, such as the media
• Research participants and their wider communities
Findings have been disseminated to them through a series of feedback sessions in-country, through bilateral meetings with key partners at national level, through the distribution of summary briefs, through steering committee meetings, as well as inviting key technical staff to participate in FEMHealth and Community of Practice conferences.
Stakeholders at regional and international level were identified as the following:
• Policymakers and advisors working in multilateral organisations
• Donor organisations
• Researchers and academics
• International NGOs
FEMHealth has disseminated findings to them through a variety of channels, including publication of articles in international journals; emailing of research summaries to a wide range of networks including the Community of Practice members and maternal health networks; writing blogs to publicise key points arising from the research (e.g. for the id21 site, HEART site, Guardian newspaper and others, ongoing), and circulation of findings within academic networks. International organisations have been engaged through CoP meetings, through involvement in the FEMHealth advisory board, and through other channels, such as virtual networks. The CoP website and FEMHealth website have been key repositories for reports, news and findings.
List of Websites:
http://www.abdn.ac.uk/femhealth/
Mr. Alec Cumming
UNIVERSITY OF ABERDEEN
Tel:+441224438107
Fax:+441224438110
E-mail:alec.cumming@abdn.ac.uk
