Voir page 405 le resume en francais. En la pagina 406 figura un resumen en espanol.
Ten years into the Safe Motherhood Initiative some 585 000 women are dying annually from obstetric complications and associated diseases. The international community continues to discuss strategies for reducing the incidence of maternal mortality from direct obstetric complications. Current models favour either a community-based approach aimed at all women, supported by a first referral unit, or a referral hospital approach with a facility upgrade for responding to unexpected complications. There is little definitive research evidence, however, to support either approach.
Proving that a strategy reduces mortality depends on the study design used to implement the intervention and on the ability to measure the desired outcome. In safe motherhood programmes targeting mortality from direct obstetric complications the necessary interventions usually involve all levels of the health system rather than a single drug or procedure, and they are commonly delivered to communities rather than individuals. These factors limit applicable study designs to descriptive designs or community randomized trials. The latter offer a higher standard of proof but the nature of safe motherhood service delivery interventions is such that to be able to generalize about results is unlikely beyond the context in which trials are conducted. This makes costs prohibitive in relation to benefits. Greater feasibility is associated with a more detailed review of epidemiological data and the building up of a coherent case study, and this is also likely to yield useful results.
Our aim is to clarify the processes involved in reducing maternal mortality by reviewing national-level data from developing countries where maternal mortality ratios have declined to about 100 per 100 000 live births or below. We have built case studies around a theoretical framework that considers where women give birth and who conducts deliveries. These factors were chosen because most maternal mortality occurs close to the time of delivery and because they capture key elements of the service organization. Such data cannot indicate a cause-and-effect relationship between programme interventions and subsequent reductions in the numbers of deaths but they do provide insights into how national programmes might have succeeded in reducing maternal mortality. Most of our information comes from internal government reports, annual progress reports of health ministries, national development plans, and discussions with health ministry officials and other key people. In order to elucidate the processes that potentially contribute to national success we drew on project data where available. The case studies were extremely useful but data with which to build on them were neither readily accessible nor comprehensive.
Our review of country programmes and projects has yielded four basic models of care, best described by organizational characteristics relating to where women give birth and who conducts deliveries. The range is from home delivery by a nonprofessional (which includes traditional birth attendants, relatives, and other community workers with brief health training) (Model 1) to delivery in a facility with comprehensive essential obstetric care by a professional (Model 4). Some features of successful models are shown in Table 1. Table 2 lists national programmes and projects that exemplify each model of care and their maternal mortality ratios.
Table 1. Models of safe motherhood care: features of successful service organization
Who delivers Where birth takes place
Home Basic essential
Non-professional Model 1:
family or provider
Professional Model 2: Model 3:
provides-basic provides basic
essential obstetric essential
care; family or obstetric care;
organizes access organizes access
to essential to essential
obstetric care; obstetric care;
essential obstetric essential obstetric
care available care available
Professional Model 4:
provides basic and
Table 2. …