Research involving pregnant women has particular ethical and practical issues. Perhaps partly because of this, much of obstetric and anaesthetic practice on the labour ward was traditionally based on little available evidence, or else little attention was paid to what evidence there was. Fortunately, there has been increasing reliance on published studies in guiding management, although in the quest for ‘evidence-based’ decisions, it is often forgotten that the best evidence there is may be far from perfect, because the ideal randomised controlled trial (RCT) has not been done. For example, the question of whether loss of resistance to air is indeed associated with a greater incidence of accidental dural tap than loss of resistance to saline would require a huge study, which would involve more centres and take more time than is practicable. In this situation one is left with methodologically weak studies (e.g. retrospective reviews), which may suggest a causal link but no more.
In the generally quoted hierarchy of evidence, the best of all is the systematic review, in which all known relevant RCTs are screened for correct methodology and the results pooled to increase power. Next, prospective RCTs themselves are still considered the gold standard for comparing different treatments or courses of management, as long as they are of adequate size. Then, in descending order, come non-randomised, single group, cohort or case-control studies; non-experimental studies; and finally case reports and ‘expert’ opinions.
Despite the fact that RCTs remain the gold standard, they are increasingly difficult to do, owing to constraints on time and funding. Many centres are now collaborating to produce multicentre databases as a way of collecting evidence pertaining to important clinical issues.
Problems and special considerations
Ethical issues are related to:
• The special vulnerability of pregnant women who are going through an intensely emotional time
• The increasing involvement of mothers in decisions affecting their pregnancy and thus their exposure to many potentially difficult choices
• The fact that many drugs in current obstetric anaesthetic practice are not licensed for use in pregnancy (largely related to the cost to manufacturers of separate trials in this group)
• The often uncertain effects of experimental drugs and procedures on the pregnancy, labour or fetus
The issue of consent maybe cause for discussion. It has been argued that a woman in labour is unable to give truly informed consent for inclusion in a study (or even for a procedure such as an epidural) because of the pain and distress she may be suffering, especially if drugs such as pethidine have been given. This makes studies of epidural techniques especially difficult, since it may not be possible to identify in advance women who might go on to request epidural analgesia. An especially controversial area concerns the calls for randomised studies of epidural versus non-epidural analgesia, in order to assess side effects in particular; some would consider it unethical to withhold the only really effective form of analgesia from mothers, whereas others claim it is unethical not to do so if it is the only way of properly evaluating the effects of epidurals.
The care of pregnant women has traditionally been something of a battleground between various medical and non-medical staff, and the risk that a well-intentioned study may be viewed as an intrusion into a normal process should not be taken lightly. Most maternity units have multidisciplinary groups or committees for the approval and oversight of proposed studies, in addition to the usual regulatory structures (research and development departments, research ethics committees, etc.), since even purely ‘anaesthetic’ studies are likely to have implications for obstetric and midwifery practice.
Practical difficulties of conducting obstetric anaesthetic studies relate to:
• Obtaining consent, discussed above (see also Chapter 169, Consent)
• Defining a homogeneous group of subjects (the ‘standard primip’ was suggested over 30 years ago as suitable for the majority of studies; women in this category are aged 25, healthy, with a full-term normal singleton pregnancy and no malposition or malpresentation)
• The fact that labouring women in particular are demanding in terms of requesting information and determining their own management
• The largely unpredictable nature of the workload
In addition, since all subjects must have the right to withdraw from any study at any time without penalty, randomised studies may suffer from considerable drop-out rates (e.g. those studies of epidural versus non-epidural analgesia). Since a mother’s situation may change suddenly during labour, this may further increase the drop-out rate. Finally, the records relating to the research may easily become lost among the voluminous paperwork that passes through the labour ward.
Management options
Apart from the considerations mentioned above, research methods are as for any other clinical situation. The practicalities of labour ward research preclude many topics from being suitable for study, even though they may be of interest and clinical value. Many potential or actual studies suffer from either too few subjects or too rare an outcome, or both; thus, for example, many complications that are relatively uncommon can only be meaningfully studied in multicentre trials (e.g. dural taps, neurological complications). On the other hand, studies comparing different regional anaesthetic techniques are relatively easy to conduct in a single unit, since so many anaesthetic interventions involve regional techniques.
As for any project, an obstetric anaesthetic RCT requires appropriate consideration of:
• The primary hypothesis (what question is the study asking - and is it worth asking?)
• The subjects to be studied (inclusion and exclusion criteria)
• The type of data collected (what is being measured and how)
• The methods of analysis (which statistical tests to use)
• The overall power of the study (related to the number of subjects per group, the size of the difference between the groups, and the statistical test used)
Finally, it should be remembered that a statistically significant result may not be clinically significant (e.g. a 30-second difference in onset of epidural block), and also that a statistically significant result found in a single study may still be a chance finding (albeit with a likelihood of less than 5% if a probability of 0.05 was taken to denote ‘statistical significance’). In addition, just because the evidence supporting a particular course of action or management may be overwhelming, it does not necessarily follow that clinicians will adhere to it.
Key points
• For research on the labour ward, basic ethical considerations apply.
• Subjects are especially vulnerable but may be especially demanding.
• Practical difficulties include obtaining consent and the unpredictability of labour.
Further reading
American College of Obstetricians and Gynecologists. Committee Opinion no. 646. Ethical considerations for including women as research participants. Obstet Gynecol 2015; 126: e100-7.
George RT, Butcher M, Yentis SM. Pregnant women’s views on informed consent for research in labour. Int J Obstet Anesth 2014; 23: 233-7.
Lupton MG, Williams DJ. The ethics of research on pregnant women: maternal consent sufficient? BJOG 2004; 111: 1307-12.