Last week, proposals in the UK to lower the 24 week deadline for abortion were rejected in the House of Commons. Decisions involving late abortion and the care of extremely preterm infants often result from strongly held beliefs rather than evidence. To a large extent this is understandable as these decisions are personal. However, when moral judgments are imposed on those holding different views and when the rights of the infant, be it to life or death, are in danger of colliding with those of parents and society the need for evidence is paramount.
In the accompanying study, Field and colleagues compare survival figures for two adjacent time periods, 1994-9 and 2000-5, for infants born at a gestational age of 22 weeks to 25 weeks and six days in a geographically defined region of the United Kingdom, the former Trent health region.1 During this 12 year period, there was no change in the number of extremely preterm infants who were alive when born but died before admission to a neonatal unit. Survival to discharge significantly improved for infants born at 24 and 25 weeks but not for those born at 23 weeks. No babies born at 22 weeks’ gestation survived in either period.
Data like these that describe temporal trends in geographically defined populations are important. Single centre studies are confounded by selection bias and tend to overestimate the likelihood of survival. The improved survival over time of more mature preterm infants shows how important it is to base decisions on as contemporaneous data as possible.
The upper gestational age limit for abortions in the UK remains at 24 weeks, although there is no age limit if the infant would be severely impaired if born alive or if the physical or mental health of the mother would be at serious risk were the pregnancy to continue. Of the 193 700 abortions that were recorded in 2006, 2% were at 20 weeks’ gestation or over.2 The UK Abortion Act 1967 was amended in 1990 to lower the age limit from 28 to 24 weeks, a change that was influenced by the Royal College of Obstetricians and Gynaecologists’ report Fetal Viability and Clinical Practice (1985), which noted significant improvements in the survival of infants born preterm.2
The discussion continues to centre on the notion of “viability.” However, other factors influence survival and health outcomes. Directives on the care that should be offered to extremely preterm babies should not be based primarily on gestational age because of biological variation and differing ethical perspectives.3
A recent study in the United States shows that outcomes can be predicted more accurately by considering four factors—sex, exposure to antenatal steroids, single or multiple birth, birth weight—in addition to gestational age.4 Improved brain imaging and other prognostic investigations enable more informed discussion between parents and clinicians about withdrawal of intensive support in the face of the prospect of major impairment, another real but often unquantified influence on survival.5 Thus, the care of the late second trimester fetus and the extremely preterm infant should not be determined solely by consideration of viability based on gestational age.
Field and colleagues’ data are in keeping with other geographically defined population studies from the UK (EPICure), Belgium (EpiBel), Denmark, Finland, France (EPIPAGE), the Netherlands, Norway, Sweden, and Australia.6 The new data are noteworthy, however, because they extend to 2005 and provide information that is among the most up to date available.
The assessment of health outcomes in large geographically defined populations has been severely limited by the costs and complexity of acquiring data. Epidemiological surveys have traditionally been conducted as single exercise research studies. They are expensive, funded for a limited period, and take a long time for data to be processed, analysed, and published.
Data for the Trent neonatal survey are collected by dedicated personnel and entered into a research register. Applying such robust but labour intensive methodology over many years is practical only when dealing with relatively small populations, in this case 16 hospitals and 55 000 births. A challenge for the coming years must be to harness emerging healthcare technologies to enable speedier entry of national health outcomes into the public domain.
In the UK this possibility is within our grasp. After a Department of Health review in 2003, neonatal services were organised into clinical networks with shared management and coordinated care pathways.7 This led to the development of electronic neonatal records that are now used by about three quarters of the 180 neonatal units in England.8 These records hold high quality standardised data on every admission to a neonatal unit; they support day to day clinical care, management, and commissioning; and they provide instant access to a full clinical record by the receiving hospital when a baby is transferred.
The potential provided by these electronic records to move the ascertainment of neonatal health outcomes out of the research arena and into routine NHS processes is being looked at by the Neonatal Data Analysis Unit.8 The unit was established to support and develop the use of electronic clinical data for audit, service evaluation, surveys, and research. It is only with information on appropriately case mix adjusted outcomes in near contemporaneous cohorts that parents, clinicians, and governments will be able to make truly informed decisions.