• Call us now: 0300 303 0444
  • Call us now: 0300 303 0444
Analysis

Look out for the limitations

6 September, 2010

Look out for the limitations

The RCM's Janine Stockdale, Mervi Jokinen and Sue Macdonald provide a brief commentary on Wax et al's meta-analysis on the safety of home birth as opposed to planned hospital birth. The RCM’s Janine Stockdale, Mervi Jokinen and Sue Macdonald provide a brief commentary on Wax et al’s meta-analysis on the safety of home birth as opposed to planned hospital birth.


Midwives magazine: September 2010



On 1 July, the American Journal of Obstetrics and Gynecology published an online meta-analysis about the safety of home birth (Wax et al, 2010). Analysing data from multiple studies from around the world, the authors concluded that – while women who had home births experienced less medical intervention – having a baby at home was associated with an increase in the neonatal mortality rate.

The authors warned that: ‘This investigation identified a doubling and tripling of the neonatal mortality overall and among non-anomalous offspring respectively’ (Wax et al, 2010). Recognising that this new evidence is at odds with previous meta-analytic reviews investigating the safety of home birth (Olsen, 1997; Olsen and Jewell, 2009), readers are encouraged to consider the limitations inherent in the Wax et al (2010) study before accepting the evidence presented: 

✲ The validity of any meta-analysis is dependent upon the careful selection of studies that are capable of answering the research question – Wax et al (2010) includes data in the analysis that was collected as far back as 1976 (Koehler et al, 1984). Home and hospital birth has significantly changed over the last 35 years. On the other hand, the author’s studies such as Johnson and Daviss (2005) that found that planned home birth was as safe for low-risk women as hospital birth when women were cared for by a registered midwife

✲ The authors fail to provide information concerning the degree of homogeneity between the studies. Study characteristics suggest that important differences exist between the included studies; for example one study reported ‘no formal policy for planned home delivery’ (Ackermann-Liebrich et al, 1996), while another investigated the ‘Home Birth Demonstration Project’ (Janssen et al, 2002)

✲ The authors do not provide information about the weighting and effect size of the individual studies. Without this information, it is impossible to tell what the contribution of each study was to the overall findings. The potential for a large study (n=665,363) to overshadow a smaller study (n=11) cannot be accounted for. The ability to generalise is therefore threatened as the meta-analysis is likely to reflect the design and participant characteristics of the larger study (Thompson and Picock, 1991)

✲ The sensitivity analysis carried out by the authors infers that neonatal deaths are not always increased in the home birth group. For example, when studies that included home births attended by someone other than certified midwife/nurse were removed, the odds ratio for neonatal deaths among all and non-anomalous newborns were no longer statistically significant.

In conclusion, although the authors state that ‘the current meta-analysis shows that planned home compared to planned hospital births are associated with significantly less maternal and newborn medical intervention and morbidity particularly among selected low-risk women, cared for by highly trained and regulated midwives who are integrated into the health care system,’ this is not the overall emphasis of the discussion. To the contrary, the commentary provided by the authors is dominated by the apparent increase in the neonatal mortality rate.

While the meta-analysis by Wax et al (2010) claims that there is a current and real risk related to planned home births, the limitations inherent in the study challenge these findings. This study by Wax et al (2010) is a demonstration of how research findings can be politicalised and easily misconstrued if midwives and policy-makers do not assess its quality sufficiently. 


References

Ackermann-Liebrich U, et al. (1996). Home versus hospital deliveries: follow-up study of matched pairs for procedure and outcome. BMJ  313: 1313-8.

Janssen PA, et al. (2002) Outcomes of planned home births versus planned hospital births after regulation of midwifery in British Columbia. CMAJ 166: 315-23.

Johnson KC, Daviss B. (2005) Outcomes of planned home births with certified professional midwives: large prospective study in North America. BMJ 330: 1416.

Koehler NU, Solomon DA, Murphy M. (1984) Outcomes of a rural Sonoma county home birth practice: 1976-1982. Birth 11: 165-9.

Olsen O, Jewell D. (2009) Home versus hospital birth (edited version – no change to conclusions). Cochrane Database of Syst Rev (1998) 3: CD000352.

Olsen O. (1997) Meta-analysis of the safety of home birth.  Birth 24(1): 4-9.

Wax JR, et al. (2010) Maternal and newborn outcomes in planned home birth vs hospital births: a meta-analysis. Am J Obstet Gynecol See: http://dx.doi.org.elib.tcd.ie/10.1016/j.ajog.2010.05.028 (accessed 2 July 2010).

Thompson SG, Pocock SJ. (1991) Can meta-analyses be trusted? The Lancet 338: 1127.






Printer-friendly version