HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH REGISTER		Author Keyword(s) Vol Page [Advanced]

To: ADC Online Letters and ADC Education and Practice Letters

Electronic Letters to:

Original articles: D J Evans and M I Levene Evidence of selection bias in preterm survival studies: a systematic review Arch. Dis. Child. Fetal Neonatal Ed. 2001; 84: F79-F84 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Selection bias in preterm survival studies

Peter Macfarlane   (12 April 2001)

Treatment decisions and survival data

Malcolm Battin   (11 May 2001)

Selection bias in preterm survival studies 12 April 2001
Peter Macfarlane, Consultant Paediatrician Department of Child Health, Rotherham General Hospital, UK Send letter to journal: Re: Selection bias in preterm survival studies peter.macfarlane{at}rgh-tr.trent.nhs.uk Peter Macfarlane	Dear Editor, The systematic review by Evans and Levine[1] identifies sources of selection bias in reporting preterm survival and recommends cohort characteristics, which should be defined to avoid bias in future studies. To these should be added the more fundamental source of confounding; that of lack of uniformity in defining a ‘live birth’, particularly around the margins of viability. Statute law does not consistently define the condition of being born alive. The UK definition of stillbirth (after 24 weeks) for the purpose of registration and ONS data (eg Section 41 of the Births and Deaths Registration Act 1953, amended by the Stillbirth Definition Act 1992), specifies a stillborn child as ‘… did not ….. breathe or show any other signs of life’. This broadly matches the WHO definition of a live birth as ‘….. breathes or shows any other evidence of life such as the beating of the heart, pulsation of the umbilical cord, or definite movement of voluntary muscles’ (International Statistical Classification of Diseases and Related Health Problems 10th revision 1992). This contrasts with a 1987 House of Lords Select Committee, being ‘born alive’ requires more than the presence of the heart beat alone and that in addition a ‘reasonable attempt to breathe’ (a matter of clinical judgement) is required. This was accepted in the context of interpreting the Infant Life Preservation Act 1929 This approach has been endorsed by the Department of Health and Coroners (personal communication). The purpose of this interpretation may well be to avoid parents having to register a non viable foetus (say at 18 weeks) who is born with some beating of the heart but no breathing as a live birth and then subsequent death. There is therefore much latitude in interpreting ‘signs of life’. A birth, especially at the margins of viability, may arbitrarily be reported as a live birth and perhaps later, a neonatal death, or a fetal death (stillbirth if after 24 weeks gestation) depending on which definition is chosen by the professionals present at the birth, exercising their clinical judgement. Clearly this will affect reporting of mortality and outcome around this gestation. References (1) Evans DJ, Levine MI. Evidence of Selection Bias in Preterm Survival Studies: a Systematic Review: Arch Dis Child Fetal Neonatal Ed 2001:84:F79-84.
Treatment decisions and survival data 11 May 2001
Malcolm Battin, Doctor National Women's Hospital , Auckland, NZ Send letter to journal: Re: Treatment decisions and survival data Malcolmb{at}ahsl.co.nz Malcolm Battin	Dear Editor, Evans and Levene have endeavoured to review the published survival data for infants born at less than 28 weeks, to identify bias and to make recommendations facilitating more accurate comparison of the published survival rates [1]. Practices regarding resuscitation vary between institutions and may change over time. In one study from a large regional centre in British Columbia, not included in this review, a cohort from 1991-3 demonstrated a substantial increase in the proportion of labour and delivery room deaths in 23-week infants compared with an earlier cohort [2]. Initial treatment decisions are clearly important for the interpretation of survival data. Thus, in addition to the proposed minimum data set, it would be useful to have an indication of whether active resuscitation was commenced or not. A possible alternative to this is reporting delivery room deaths and deaths in the neonatal unit separately but this is not perfect as resuscitation may have been initiated but failed. (1) Evans DJ, Levene MI. Evidence of selection bias in preterm survival studies: a systematic review. Arch Dis Child Fetal Neonatal Ed. 2001;84:F79-84. (2) M Battin, EW Ling, MF Whitfield, SB Effer. Has the outcome for extremely low gestational age (ELGA) infants improved following recent advances in neonatal intensive care? Am J Perinatol. 1998;15:469-77.