Collaborative Birth Audit, 1998. Comparing waterbirth with dry birth

Baum (1990) in a perspicacious review of the ethics of clinical research points out the impossible dilemma facing clinical researchers:

'... If they rigorously attempt comparative tests of new and existing therapies on patient-subjects they will be accused of using the sick to gain knowledge. If they apply an unvalidated therapy to patients in general without previous controlled trials, they will be accused of taking risks with people's health....'

He goes on to point out that '...much, if not most, of contemporary clinical practice is essentially based on forms of guesswork. Many acceptable routine forms of treatment have not been properly scientifically validated...' Baum believes that the time has arrived when the guessing has to stop.

Waterbirth in the UK began in the 1980s as an unvalidated therapy. The dilemma elucidated by Baum is modified by the fact that waterbirth studies do not use the sick as instruments to gain knowledge. Only low risk mothers are ever studied. They are not generally sick, although it is, of course, possible for the health status of the mother and/or the baby to become compromised during labour and delivery. Usually, though, this would result in the mother leaving the water for assisted delivery in bed or operating room.

With regard to the other horn of the dilemma, there have been a number of calls for randomised controlled trials to answer questions about the safety of waterbirth (e.g. Atalla and Weaver, 1995; Kitzinger, 1995a,b, Nikodem, 1999). There are those (Garland and Jones, 1994, 1997) who have reservations about such trials. As the author of this report is the Jones in 'Garland and Jones', it is important to be clear on the reasoning behind these reservations.

The first issue centres on the fact that the modern Western lifestyle means that Western women have relatively few opportunities to give birth. This being so, it seems unreasonable to randomly assign them to an arm of a trial in which the mode of delivery is inconsistent with their wishes (i.e. allocating them to the mode of delivery that they least prefer).

The second issue concerns the question of non-compliance by the mother if she is asked to deviate from her preferred birth plan. The subject is unresearched but the suspicion exists that non-compliance may arise even if the mother has consented to deviate from her preferred birth plan (e.g. has consented to join a randomised, controlled trial). This may happen because the mother experiences an altered state of consciousness during labour. Extension of this line of thought suggests the possibility that psychosomatic responses may, in extreme cases, impair the progress of labour.

Nevertheless, the need to promote evidence-based midwifery practice has driven certain maternity departments to monitor the respective outcomes of the two types of delivery. Some of these studies have been reported in the British professional literature (Burns and Greenish, 1993; Garland and Jones, 1994, 1997; Nightingale, 1994, 1995; Burke and Kilfoyle, 1996).

However, Garland and Jones (1997) point out that it is impossible to generalise these findings unless adequate multi-centre data becomes available for analysis. At present, relatively minor differences in data definition and collection impede full comparison between the existing published papers. Alderdice, Renfrew and Marchant et al (1995) revealed that many maternity units did not collect comparative data or, indeed, any at all.

To address this problem a five-centre pilot audit was carried out from January 1st, 1998 through December 31st, 1998 at the Birth Units of the following Hospitals:

Wlingdon Hospital, Uxbridge, West London

The Maidstone Hospital, Maidstone, Kent

Northampton General Hospital, Northampton

The Hospital of St. John and St. Elizabeth, St John's Wood, North London

Torbay General Hospital, Torquay, South Devon