Printed in Great Britain

A total of 542 women aged 15 to 44 years died during the 10-year period 1976 to 1985 in the control area of Matlab, an area with a population of 90 000, representative of many other rural areas of southern Bangladesh. The corres-ponding age-specific mortality rate was 290 per 100 000 women 15-44 years. These deaths have been analysed retrospectively, using information collected through the Demographic Surveillance System set up by the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR.B) and verbal autopsies conducted in the homes. Of these deaths, 175 (32%) were due to infectious diseases, 163 (30%) to direct obstetric complications, 67 (12%) to injuries, and the remaining 26 % to other causes. Cause-specific and proportionate mortality rates showed a positive association with age for deaths due to infectious diseases, non-infectious diseases and unspecified causes, and an inverse association with age for deaths due to injuries. These rates showed a peak in the intermediate age group 25 to 34 years for deaths due to direct obstetric causes. No consistent trends were visible when annual rates were studied over time. Prior to death, 42% of the women were attended by traditional practitioners, and 33 % were not attended at all. Demographic impact is discussed, emphasizing the contribution of obstetric causes to overall mortality. Priorities for health policy implications are proposed, focusing upon a strong maternity care programme, and improved availability of female health personnel, in the context of the socio-cultural constraints imposed on women in poo