Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
The purpose of this article is to review current strategies for the reduction of maternal mortality and the evidence pertinent to these strategies. Historical, contextual and current literature were examined to identify the evidence base upon which recommendations on current strategies to reduce maternal mortality are made. Current safe motherhood strategies are designed based mostly on low grade evidence which is historical and observational, as well as on experience and a process of deductive reasoning. Safe motherhood strategies are complex public health approaches which are different from single clinical interventions. The approach to evidence used for clinical decision making needs to be reconsidered to fit with the practicalities of research on the effectiveness of strategies at the population level. It is unlikely that any single strategy will be optimal for different situations. Strengthening of the knowledge base on the effectiveness of public health strategies to reduce maternal mortality is urgently required but will need concerted action and international commitment.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The continuing high maternal mortality in low resource countries is evidence that there is a need to identify and implement those strategies that are most effective at reducing maternal mortality. A strategy can be defined as a set of plans and courses of action, which together will result in the achievement of a goal such as reduction of maternal mortality. We thus use the term strategies to refer to high level, complex public health goals that underpin programmatic and policy level initiatives. Single clinical interventions do not count as strategies but are likely to be an integral part of strategies implemented at health facilities. Thus, comprehensive emergency obstetric care (EmOC) is a strategy, while carrying out a caesarean section or giving blood transfusion are specific clinical interventions. Strategies are not only of interest to policy makers. Scientists are needed to provide the necessary evidence of effectiveness, planners to convert them to programmes and programme managers and health care providers to implement them.
This review of strategies is aimed at all of these groups. Some strategies that may reduce maternal mortality are not specific to maternal health and we have included a brief description of these health system and contextual characteristics, as their importance should be underscored. However, our main focus in this review is on safe motherhood strategies where attention is placed on actions directly pertaining to maternity care.
Many public health strategies designed to reduce maternal mortality were based more on deductive reasoning (or on fashion) rather than evidence. The need for decisions about all forms of health care to be informed by the best available evidence is now generally accepted. Although the evidence base for recommending one strategy over another is generally lacking,1 in commenting on strategies for the reduction of maternal mortality we have where possible referred to the evidence that supports them. We have also commented on the nature of evidence in relation to safe motherhood strategies and the work that remains to be done before strategies can be based on credible evaluation research about their effectiveness.
The experience in countries such as Sweden, Sri Lanka and Malaysia that were historically more successful in reducing maternal mortality than other comparable countries is a source of insight into successful strategies. The gathering and analysis of information in the form of birth and death registration is one of the earliest recorded successful strategies. The employment of well-trained midwives working in a properly regulated environment together with free and accessible health care has also been considered successful.2–4 The value of modern hospital technologies is well documented but it is instructive to note that Sri Lanka and Malaysia were able to benefit from these technologies only because they had a functioning health system to deliver them.2,3 The question of whether social and economic development will lead to a reduction of maternal mortality has been postulated as well; however, data (Fig. 1)5,6 suggest that even poorer countries like Sri Lanka can achieve relatively low maternal mortality ratios.
These historical lessons cannot provide clear evidence regarding the adoption of strategies in individual low resource countries. The historical evidence afforded is contextual and the determinants of maternal mortality are complex and vary from place to place. Furthermore, the data provide no evidence of cause and effect. Strategies like vital registration require functioning systems at both government and community level. Finally, some strategies are too expensive for some countries.
The health sector and wider context
A country's socio-economic, cultural, political and religious context influence the health of the population.7 Characteristics that specifically affect maternal health include inequality in terms of child-rearing and decision-making authority and inaccessible safe abortion. Such factors have been addressed using a human rights approach to lobby for socially or legally instigated social change.8 Policy makers can use knowledge of these factors to benefit women and mothers. A situation analysis conducted to inform policy making can enhance lobbying power by highlighting the impact of such contextual factors. Policy makers may improve the targeting of strategies by demonstrating that cost effectiveness may be maximised by focussing on the most disadvantaged groups.9
In addition to these wider contextual issues, a properly running health system is necessary for the implementation of strategies to reduce maternal mortality.10 Apart from the essentials of providing basic equipment, supplies and referral systems, financing and human resource organisations are equally important components of the health system. User charges are found to be regressive in many countries, favouring the better-off, reducing utilisation by the poor and leading to destitution and the selling of assets. The solution remains elusive and further research is needed to assess the various strategies currently being employed in low income countries.11
Of the current major public health issues in low resource countries, safe motherhood is unique in requiring large numbers of clinical staff including some trained in surgical techniques. In comparison, programmes for malaria control, HIV/AIDS and tuberculosis have a greater capacity to be effective with a high proportion of non-clinical staff. A human resource strategy with the objective of ensuring a supply of appropriately trained staff is thus essential. Decisions may be needed at any of three levels. Firstly, many developing countries now face problems of retention of trained staff due to migration across countries and also from rural to urban areas. Secondly, in some countries with a small number of doctors and midwives it may be more appropriate to train all health professionals to competence level in obstetric care for complications. In other countries with a larger complement of health professionals, specialist training may be more feasible. Thirdly, debate has focussed on which cadre can or should be permitted to perform specialised obstetric procedures by delegation. The choice could lie between specialists or generalist doctors or clinical officers/medical assistants/nurse-midwife practitioners. Although delegation has been shown to be a safe and appropriate option,12 the decision is very much one to be decided on the basis of finding an optimal balance between quality of care and coverage. Similar arguments have been raised in relation to the training of anaesthetists. For patients anaesthetised by specialist anesthesiologists and by non-anaesthetist grades (such as doctors with anaesthesia training and nurse-anesthetists), modelling suggests that the improved coverage provided by trained non-anaesthetist grades would result in much higher survival even when calculated on the basis of an unrealistically low survival rate.13 Professional associations have potential influence in all these examples of human resources strategies.
Safe motherhood strategies
A review of the safe motherhood literature demonstrates that three main approaches to the design and planning of strategies are used. The earlier approaches date from the global call for action in safe motherhood in Nairobi in 1987. In the decade that followed, safe motherhood strategies were developed based on the different phases in a woman's reproductive cycle. These phases were classified into pre-pregnancy, antenatal, delivery and the postpartum period. Examples of these strategies include the World Bank's Safe Motherhood Initiative14 and the World Health Organization's Mother Baby Package.15
At around the same time, a number of strategies used a framework based on primary health care principles where primary prevention (preventing a condition from occurring through education and services) was distinguished from secondary (detection and treatment of conditions early) and tertiary (treatment of conditions to reduce case fatality) prevention.16 Examples include MotherCare's Pathway to Survival17 and the Three Delays Model for EmOC.18
More recently, health system factors began to influence the development of strategies. Approaches recommending intervention through the health sector and contextual actions, such as rights, societal actions and legislation, became prominent. Examples of these include the Safe Motherhood's Inter-Agency Group's Action Agenda,19 UNICEF's Programming for Safe Motherhood20 and the World Health Organization's Making Pregnancy Safer (MPS) initiative.21
The range of strategies described demonstrates the degree of complexity and overlap between different strategies. Many earlier approaches remain influential today. It would be difficult to comprehensively review all current strategies, so we have chosen to highlight a selection according to three common areas of debate: address demand or improve supply; seek preventive approaches or treat complications; and the question of whether strategies should be focussed or broad based.
Address demand or improve supply?
Much emphasis has been placed in medical education on the need for community orientation of the undergraduate curriculum.22 The approach has also been described as one that should be promoted by professional associations.23 Safe motherhood strategies aimed at identifying needs, and addressing the demand for care are sometimes described as ‘community-based strategies’ and include community mobilisation and training of traditional birth attendants (TBAs). Strategies to improve supply are generally directed toward interventions that improve the availability and quality of services in health facilities and through the formal health sector. The distinction between the two extremes can become blurred when considering specific approaches. For example, the Indonesian government's efforts to place qualified midwives in villages may be considered a supply side strategy, as the midwives are within the formal health sector. However, because the midwives practice in villages, they are based within the community, and by bringing services closer to the community, the strategy aims to improve demand. We thus believe that strategies relating to supply and demand are best considered along a continuum, particularly as supply and demand factors interact closely and changes in one can often influence the other.
Interest in community-based strategies has recently been boosted by evidence pointing toward the effectiveness of community-based participatory interventions in Nepal.24 In this cluster randomised controlled trial, a female facilitator (non-health professional) convened women's group meetings monthly to raise awareness around childbirth. Although the study was not designed to measure maternal mortality, intervention clusters showed a maternal mortality ratio 80% lower (2 deaths compared with 11) than control clusters and reduced neonatal mortality by 30%.
The training of TBAs is another strategy upon which much emphasis has been placed. In many countries women prefer TBAs to midwives as their delivery attendant. TBAs are also likely to remain as delivery care attendants for some time because of difficulties experienced in posting trained professionals to rural areas in many developing countries. A recent systematic review showed that TBA training appears to increase antenatal care attendance rates by 38%,25 but no improvements in maternal mortality could be found.26 The evidence in support of TBA training remains inconclusive although it is currently recommended that support of existing TBA programmes should be done selectively while giving precedence to other programme options, which are based on stronger evidence of effectiveness.27
Service quality improvements
In relation to maternal deaths the gathering of information on deaths with a view to finding out why the deaths occur, and what can be done to prevent them, is the keystone of quality assurance strategies. This can take the form of verbal autopsies in the community, facility-based maternal death reviews, confidential enquiries, reviewing cases of severe maternal morbidity (near-misses) and criterion-based clinical audit of life threatening complications. These methodologies are strongly supported by expert opinion and have received a recent endorsement from the World Health Organization.28 Many evaluations of audit have been conducted showing some benefit but absolute scientific proof that it is worthwhile is not available. A Cochrane review of randomised controlled trials of audit and feedback showed them to have a better effect on health care practices and health care outcomes than other strategies although the effect size could be modest.29 Despite this, in a review of the evidence for the role of audits to improve the quality of obstetric care, it was concluded30 that enquiry into the quality of obstetric care is important and that this may be particularly so for facilities in developing countries where there is evidence that the services fall short of acceptable standards.
Seek prevention or treat complications?
Primary prevention of maternal mortality is exemplified in the consideration of family planning as a strategy. During the 1980s, family planning was presented as one of the key strategies for maternal mortality reduction in developing countries.31,32 If accepted by a large proportion of the population, and if used continuously for prolonged periods, contraceptive methods should, at least in theory, contribute to lowering the high levels of maternal mortality. Family planning may prevent unwanted pregnancy (and illegal abortion), redistribute births from high to low risk categories, reduce the total numbers of births and have direct benefits from the contraceptive methods themselves.33 Yet various reports examining the potential impact of family planning on the reduction of maternal mortality have suggested disappointing effects.33–35 There is no doubt that widespread use of contraceptives will reduce the total numbers of maternal deaths hence lower the maternal mortality rate, as fewer women will be exposed to the risks of pregnancy. However, the effects on the maternal mortality ratio, that is the risk of death once a woman is pregnant, are not so clear.33,34,36 The vastly lower mortality ratios in the developed world when compared with developing countries cannot be attributed to changes in the demographic distribution of births.33 A study in Bangladesh has also convincingly shown that while increased use of contraceptives was associated with a steady decline in the maternal mortality rate, no such effects were observed for the maternal mortality ratio.37,38
Another preventive approach is being advocated for chronically malnourished populations in the form of micro-nutrient supplementation, which appears attractive as a potential intervention to reduce maternal and fetal mortality because it is believed to be cheap, safe and easier than the more fundamental changes in society that may be required. Widespread appeals for the promotion of micro-nutrient supplementation of pregnant or reproductive age women have been made, and some agencies have incorporated supplementation strategies in their policy agenda.39,40 Very few studies have been able to explore direct links between nutritional supplementation and maternal mortality or severe morbidity, however, and the evidence base is weak. An exception to this may be vitamin A and calcium supplementation in high risk women. Vitamin A and its precursors may affect maternal health through improvements of the immune and haematological status of the pregnant woman. A large randomised double blind placebo controlled trial of vitamin A and β-carotene in Nepal suggested that vitamin A and/or β-carotene may be associated with a 40% reduction in maternal mortality.41 Much further work is needed before such supplementation can be proposed as a strategy for reducing maternal mortality.42 Reviews of calcium supplementation trials during pregnancy provide strong support for the supplementation of pregnant women with calcium as a means of preventing pregnancy-induced hypertension and pre-eclampsia in communities with low calcium intake.43 Two meta-analyses of randomised controlled trials have consistently shown a protective effect of calcium supplementation on the incidence of pre-eclampsia.44,45 The Cochrane review of randomised controlled trials concludes that sufficient evidence is available to support calcium supplementation for pregnant women at high risk of pregnancy-induced hypertension and in communities with low dietary calcium intake.43 There is little doubt that iron supplementation in pregnancy of around 100 mg elemental iron daily improves maternal iron status and haemoglobin levels during pregnancy and immediately after delivery in both industrialised and developing countries.46,47 Although trials have demonstrated an effect on iron status and haemoglobin levels, there has been no proof of effect on actual health outcomes.
The rationale for the widespread introduction of antenatal care (ANC) has been the belief that early signs of, or risk factors for, morbidity and mortality can be detected and that effective interventions are possible.48 Reviews examining the effectiveness of formal risk assessment in pregnancy have however concluded that the risk approach may be neither effective in preventing maternal death nor in ensuring rational use of resources.19,48 Evaluations of the performance of risk scoring systems in developing countries have also shown that complications such as dystocia and postpartum haemorrhage cannot be adequately predicted.49–51 This has led to a shift in the emphasis of safe motherhood strategies from universal ANC to one of universal access to professional delivery care.19 This should not lead us however to overlook the value of ANC in the detection and treatment of pregnancy-related complications. Although there is a lack of strong evidence on the effectiveness of the content, frequency and timing of visits in ANC programmes, ANC offers an opportunity for alerting the woman to the risks associated with the pregnancy and for discussing and planning her options for professional care during delivery. There is also evidence that this can be achieved with fewer than the standard number of visits.52 Women seeking antenatal care may be more likely to seek professional care during delivery.51 ANC therefore still has importance as a potentially effective instrument to ensure better use of obstetric services.
Skilled attendance at delivery
Increasing the proportion of deliveries with skilled attendance is regarded as a crucial intervention strategy and is widely advocated by international agencies.53 It is a complex concept, which has been depicted by the schematic framework in Fig. 2.54,55 At first sight it may appear that skilled attendance incorporates all that is needed to prevent maternal deaths. It has a preventive component of ‘watchful expectancy’ for normal deliveries as well as referral to professional care for emergencies. Furthermore, the needs and demands of women and their carers in the community can be addressed within this concept, although it is often forgotten that skilled care should also include women's perspectives.56 Indeed it has been calculated that this model of care could prevent between 16% and 33% of maternal deaths.54
Emergency obstetric care
EmOC is a well-known concept. It is a package of interventions focussed on the direct obstetric complications that cause the majority of maternal deaths.57 A recent comprehensive review58 demonstrated the effectiveness of EmOC by a combination of quasi-experimental, observational, historical and ecological studies plus analysis of trends in MMR in relation to introduction of EmOC. This raised the level of evidence supporting the strategy from one of clinical observation. EmOC is an integral component of skilled attendance but the extent to which the two coexist and function varies. Because resources are limited it is important to know their relative effectiveness in reducing maternal mortality. The problem here is that this must depend on how women access EmOC. If they only access it as a result of utilising care for normal delivery, then the EmOC service would be underutilised without a parallel functioning normal delivery service. Studies38 do suggest that EmOC can be effective in the absence of delivery by a skilled attendant. Historical studies of the development of services in Europe and the United States tend to support this but also show that the two strategies are synergistic.
Make strategies focussed or take a broad-based approach?
Two strategies, post-abortion and postpartum care, are used here as examples of focussed or narrow strategies that target specific periods of high risk during pregnancy. The resources required to implement targeted strategies are likely to be more feasible to mobilise, in contrast with the very broad and currently high profile World Health Organization MPS strategy.
Each year about 70,000 women die as a consequence of unsafe abortion. Good post-abortion care can make a contribution to reducing them. The introduction of post-abortion care as a strategy has depended to a large extent on the work of the Post-abortion Consortium comprised of several agencies that work to inform the reproductive health community about possible complications related to miscarriage and incomplete abortion. Essentially the strategy consists of the scaling up of good quality post-abortion care including the use of manual vacuum aspiration instead of dilatation and curettage. The scientific evidence for its effectiveness does not extend to its effect on maternal mortality but evaluations have shown that scaling up leads to better patient care, shorter hospital stays, lower costs and increased contraceptive use and the adoption of local anaesthesia in lieu of general anaesthesia.59 The paucity of deaths from abortion in countries where this standard of care is the norm is evidence of its value.
More than 60% of maternal deaths occur in the postnatal period60 and a survey of women delivering in rural homes identified a 43% rate of postpartum morbidity.61 How the health services can best respond to this is still uncertain. Most postpartum deaths occur the first day after birth and their management falls within the skilled attendance or emergency care strategies. Postpartum home visits have been suggested61 but said to be a subject that needs further research by others.60
Making pregnancy safer
The World Health Organization's MPS initiative21 is an example of a very recently developed approach influenced by the evidence currently available. It focuses on health system improvement but pays attention to most aspects of safe motherhood. MPS is 1 of 11 World Health Organization priorities, an expression of its commitment to maternal and newborn health. It advocates widespread skilled attendance by midwives and health professionals throughout pregnancy, delivery and the immediate postpartum period, backed up by accessible and functioning emergency care. MPS approaches safe motherhood from a human rights perspective and has a particular focus on equity. Through its global and national advocacy activities, it seeks to keep safe motherhood as a high priority on the health and development agenda and to promote ethical, evidence-based, cost-maximising strategies. It fosters international and national intersectoral partnerships and emphasises the importance of monitoring and evaluation to generate evidence. The World Health Organization offers governments technical support to develop a national strategy and a supportive policy and legal structure. It assists in efforts to identify, apply and co-ordinate cost-effective interventions and to integrate these with other interventions like HIV and malaria both in the health sector and beyond, for example, through national poverty reduction strategies.
The need for evidence
It might be thought that a review of strategies for the reduction of maternal mortality would be based on good quality evidence for or against specific strategies. For a number of reasons this cannot be so. Even though in some instances a safe motherhood strategy may consist of a number of effective clinical interventions it is not always appropriate to utilise the individual evidence for application at the population level. Thus, although advice on the basis of risk screening may be of value in an individual case, risk screening may not be an effective public health strategy. This is because a greater number of obstetric complications actually occur in low risk women because their numbers in the population are so much greater. Additionally, single interventions need to be supplied through a functioning health system that is itself comprised of other untested interventions. Thus, although performing caesarean section to manage obstructed labour may be self-evidently beneficial, the only way of evaluating a strategy to provide EmOC is to evaluate it in the context of the health system. Further, in reaching decisions about employing a strategy, regardless of whether or not there is good evidence, it is vital to remember that the value of a strategy is very dependent on the situation in which it is to be employed. This has lead to the concept of context-based evidence-based decision making for health policies in which context plays a role in the introduction, interpretation and application of evidence.62
The quantity of evidence on the effectiveness of safe motherhood strategies is as disappointing as the quality of what is available. This is largely the consequence of a measurement gap, due to the weak health information systems in most low resource countries resulting in inadequate and unreliable data, and additionally a general lack of information on gender, poverty and technological disparities.1 Most safe motherhood strategies currently in use are based on observational evidence from surveys, demographic surveillance, case studies and qualitative studies. There are however significant problems over attribution in this type of research. It is very difficult to control confounders in real world settings, and thus to know with certainty that a particular intervention strategy caused a particular health change. For these reasons we would argue that the grades and views of evidence used for clinical decision making need to be reconsidered to fit with the practicalities of research on the effectiveness of strategies at the population level.
The generation of evidence on safe motherhood strategies must go beyond the strengthening of the evidence base towards translating this evidence into the practice of decision making. This is not just a matter for policy makers. Many others are involved and good communication between them is essential if the process of research, strategy development and decision making to programme development and implementation is to work efficiently.Figure 3 demonstrates an unsatisfactory communication network for this purpose and Fig. 4 an improved model we propose, which demonstrates the following:
•the need for the end-users of the research, the policy makers, to be involved in its conception and conduct;
•the need for the research to be contextually based and how both programme managers and those involved in the implementation of strategies have a role in providing contextual information by the collection of data or research;
•the importance of researchers disseminating their findings effectively both to policy makers and programme planners; and
•that policy makers have a need to ‘sell’ the strategies they have adopted to the doctors, midwives and field staff who are implementing them to ensure that everyone is contributing to the strategy and not running counter to it.
Despite an international consensus that identified maternal mortality as an urgent public health priority in 1987, the international commitment to address it remains inadequate. While a global safe motherhood partnership exists, it is relatively small if one compares it to the scale of mobilisation that occurred to tackle the HIV epidemic. For safe motherhood, we need a similar international commitment from the United Nations system, international donor agencies, national governments and civil society. Together they must establish a strong body of governance and mobilise resources so that safe motherhood programmes can be scaled up using known evidence-based intervention strategies. Implementation of the United Nations millennium development goals provides a unique opportunity to develop this strong, active and global partnership for safe motherhood.
This work was undertaken as part of an international research programme—Initiative for Maternal Mortality Programme Assessment (IMMPACT). See: http://www.abdn.ac.uk/immpact, funded by the Bill & Melinda Gates Foundation, the Department for International Development, the European Commission and USAID. The funders have no responsibility for the information provided or views expressed in this paper. The views expressed herein are solely those of the authors.
The authors would like to thank Dr Margaret Armar-Klemesu's contribution towards framing the outline of the paper and commenting on drafts. Lesley Milne's review of global safe motherhood strategies and Lucia D'Ambruoso's MSc dissertation database were used in writing this paper.
Competing interest statement
All authors are fully or partially employed by the IMMPACT research programme.