Objective: To review the reporting of randomised controlled trials for intrapartum interventions.
Methods: This was a literature search addressing the period from the publication of the CONSORT statement i.e. 1996-2002. Databases searched: Medline and CINAHL. Search terms: pregnancy, low-risk, normal, intrapartum, labour and labor.
Inclusion criteria: randomised controlled trials reported in English which involved women experiencing: normal or 'low risk' pregnancy; intrapartum interventions; women who had a spontaneous onset of labour.
Data extraction: timing of consent and randomisation, proportion of eligible women recruited, 'losses' to the study. Included Studies fourteen published and one unpublished study that fitted the inclusion criteria were identified. Three studies were subsequently excluded because of a lack of information about the method and a further study was excluded due to its small sample size.
Findings: Three strategies for recruitment and randomisation for intrapartum studies were reported in the literature: antenatal recruitment and randomisation; consent and randomisation conducted on admission in labour or at the time of the intervention; staged recruitment and randomisation, which may be two- or three-staged. Different study designs have done much to improve the appearance of intrapartum studies and reduce post-randomisation losses. However, the problem of bias in RCTs is not limited to attrition alone and these designs may simply be moving the problem to an earlier stage in the study resulting in selection bias and limiting generalisability.
Conclusions: The importance of accurate documentation of numbers at all stages in the research process is highlighted, in particular the number of eligible people not recruited, to allow the reader to assess the generalisability of the study.