Findings of meta-analysis cannot be relied on. BMJ 2010;341:c4033  

 

 Findings of meta-analysis cannot be relied on. BMJ 2010;341:c4033  

  1. Gill Gyte, [log in to unmask]  research associate1,
  2. Miranda Dodwell, editor, BirthChoiceUK2,
  3. Mary Newburn, head, research and information3,
  4. Jane Sandall, professor of women’s health4,
  5. Alison Macfarlane, professor of perinatal health5,
  6. Susan Bewley, consultant obstetrician/maternal-fetal medicine6

Author Affiliations:

1.      1Cochrane Pregnancy and Childbirth Group, Division of Perinatal and Reproductive Medicine, University of Liverpool, Liverpool Women’s NHS Foundation Trust, Liverpool L8 7SS
2.      2c/o NCT, Alexandra House, London W3 6NH
3.      3NCT, Alexandra House, London W3 6NH
4.      4Department of Primary Care and Public Health, King’s College London, London SE1 3QD
5.      5Department of Midwifery and Child Health, City University London, London E1 2EA
6.      6Women’s Services, Guy’s and St Thomas’ NHS Foundation Trust, London SE1 7EH

Mayor reports that a meta-analysis has linked planned home births with a twofold higher rate of neonatal mortality compared with hospital births.1 2 Closer inspection calls this finding into question.

The quality of studies in any meta-analysis is critical, but no assessment was reported. Studies were observational with many not matched adequately for confounders.

Neonatal mortality came mainly from small studies, with most weight from one larger retrospective study on birth registry data for Washington State.3 Unplanned home births are more likely to have poor outcomes, and some may have been misclassified as planned home births because birth certificates did not distinguish between them.

Differences arising from comparatively small numbers should be interpreted with caution. Differences in neonatal mortality were based on 32 deaths in 16 500 planned home births and 32 in 33 302 hospital births.2 This lacks the power recommended by the GRADE quality assessment tool (being phased in by the National Institute for Health and Clinical Excellence),4 which suggests that 200-400 events are needed. In contrast, perinatal mortality was based on 229 deaths among 331 666 planned home births and 140 among 175 443 hospital births, with no significant difference.2

Unfortunately the meta-analysis and the BMJ focused on the neonatal mortality findings. Outcomes given less prominence were no significant differences in perinatal mortality and neonatal deaths with planned home births attended by certified midwives. Mothers planning a home birth were less likely to have a preterm or low birthweight baby. All the outcomes should be viewed within the overall poor quality of the meta-analysis. Professional journals should be reporting findings in a balanced way, highlighting methodological limitations.

 

References

  1. Mayor S. Study shows higher rates of neonatal mortality with planned home births. BMJ2010;341:c3551. (2 July.)

 

  1. Wax JR, Lucas FL, Lamont M, Pinette MG, Cartin A, Blackstone J. Maternal and newborn outcomes in planned home birth vs planned hospital births: a metaanalysis. Am J Obstet Gynecol (in press). Available online 2 July 2010.
  1. Pang JWY, Heffelfinger JD, Huang GJ, Benedetti TJ, Weiss NS. Outcomes of planned home births in Washington State: 1989-1996. Obstet Gynecol2002;100:253-9.

 

  1. GRADE Working Group. Grading quality of evidence and strength of recommendations. BMJ2004;328:1490-4.

 


------------------------------------------------------------

List Master: Michael C. Klein [log in to unmask]
List address: [log in to unmask]
Web Resources: www.cfpc.ca/MCDGResource
Subscribe: [log in to unmask]&body=Si%0d%0agn%20me%20up!">Click here to Signup
Unsubscribe: Click here to Signoff List