The value of a prospective report on maternal deaths: a Trinidad experience letter

ABSTRACT

Most official measures of maternal mortality are underestimates and in 1984 it was noted that there was routine reporting of maternal deaths in only 35 percent of the world's population. Under-reporting in the USA ranges from 25-70 percent and a recent enquiry in France has estimated a 56 percent under-reporting. The problem is likely to be even worse in developing countries. Reporting of maternal deaths has been largely retrospective and Royston and Abou Zahr have identified poor record keeping, tampering and misplacement of patients' notes as factors which adversely affect the proper audit of amaternal deaths. In addition, in certain instances, deaths on the gynaecological wards, e.g. associated with ectopic pregnancy and septic abortion may not be recorded as maternal deaths, and the same can be said for maternal deaths on general medical wards. Over the 18 month period 7 May 1991 to 31 October 1992, we concluded a 'prospective' audit of maternal deaths at the San Fernando General Hospital, Trinidad. The mechanism of this audit was that the entire hospital staff was alerted and a specially designed form circulated so that complete data were entered as soon as a death occurred. There were 14 deaths giving a maternal mortality rate of 157/100 000 births. Six women died from haemorrhage (42.9 percent), with criminal abortion and hypertensive disorders of pregnancy accounting for three each. One died from puerperal sepsis and no definite cause was found at postmortem for one who died 24 h following elective Caesarean section. Our maternal mortality rate was found to be 20.7 times greater than that recorded in England and Wales in the triennium 1985-1987. This disparity also extended to the causes of deaths since deaths in Trinidad now follow a pattern that was present in the UK in the early part of this century. This rate of 157/100 000 birth represents as closely as possible the precise rate of maternal deaths in southern Trinidad. In contrast, a recent report has indicated the maternal mortality rate in Trinidad to be 30/100 000 births, a figure which compares favourably with those of other developing countries. However, this report emanated from another institution in Trinidad where deaths occurring over a 10-year period were analysed retrospectively, a method which lends itself to under reporting. Maternal deaths occurring in the community are received by the District Medical Officer who issues the death certificate. These are recorded at the Central Statistical office and only the principal cause of death, e.g. cerebrovascular accident is noted. These deaths are not recorded as maternal mortalities. Our finding in contrast indicates that using maternal mortality as an indicator, maternal care in Trinidad is several decades behind that of the developed world and, therefore, demonstrates the value of prospective data collection on maternal deaths (AU)