RAPID ESTIMATIONS OF MATERNAL MORTALITY IN COUNTRIES WITH DEFECTIVE DATA - AN APPLICATION TO BAMAKO (1974-85) AND OTHER DEVELOPING-COUNTRIES

Maternal mortality is not easy to measure directly in countries in which it remains a public-health problem. Registers in which the cause of death is reported provide a useful tool, if their coverage can be assessed. In Bamako, they showed an average maternal mortality of about 300 deaths per 100,000 live births, characterized by a pronounced U-shaped distribution: the risk of dying was largest for mothers below 20 years and above 35 years. It was possible to divide maternal deaths into direct and indirect obstetric causes, the former being predominant at all ages, except the youngest. However, such registers are not always available. Two indirect methods have been devised to assess maternal mortality from general life tables, in the absence of any information on cause of death. The first method, in which ratios of age-specific death rates for women to those for men is used, ascribes the deviation of this ratio observed at childbearing ages to maternal mortality. The second method which is based on the fact that the trend in the logarithm of women’s age-specific death rates has two different slopes, higher above the age of 50 and lower below that age, also attributes the deviation observed at childbearing ages to maternal mortality. When applied to Bamako, these two methods give results rather close to those obtained directly from the registers. When applied to a set of life tables, they provide age patterns and trends of maternal mortality for 20 countries, that seem fairly consistent. In countries in which fertility is known, they also make it possible to assess the respective contribution of fertility and obstetric risk in an observed trend. © 1990 Taylor & Francis Group, LLC.