Given the limitations inherent in medical knowledge and also the concerns about litigation [25] it is not surprising that doctors have, to some measure, come to endorse patients’ preferences and valuation. This is perhaps more commonly integrated within the taxonomy of ‘medically indicated’ interventions than is acknowledged. In fact, patient preference plays—as it ought to do—a distinctly decisive role in a large number of procedures. In gynaecological practice, this includes procedures performed for abortion, sterilisation, fertility treatment, hysterectomy for non-malignant indications or operations for prolapse: in short, in most elective surgery. I say to some measure because patients undergoing these procedures are still required to fit within a medically defined framework such as the requirement to try other forms of therapy or to reach a certain threshold of eligibility. So why should a request for caesarean section cause so much disquiet? The European multi-centre study (EUROBS) compared the attitudes of obstetricians from eight European countries, France, Germany, Italy, Luxembourg, the Netherlands, Spain, Sweden and the UK, to CSMR [25]. The clinical case description was of a 25-year-old woman who started labour at 39 completed weeks. The foetus was normal and in cephalic presentation. She insisted on a caesarean section despite being informed that a vaginal delivery was indicated, and of the higher morbidity and mortality associated with caesarean delivery. Compliance with this woman’s request for caesarean section simply because this ‘was her choice’ was lowest amongst responders from Spain (15%), France (19%) and the Netherlands (22%), and was highest in the UK (79%) and Germany (75%). Respect for patient’s autonomy was the most frequently reported justification for accepting CSMR. Fear of litigation and working in a university-affiliated hospital were associated with physicians’ likelihood to agree to patient’s request whilst female doctors who themselves had children were less likely to agree. Whilst this indicates a high level of acceptance of CSMR, it also indicated that opinion remains divided. The country differences may indicate differences in prevailing attitudes or cultures.

The perception of an intervention as being ‘indicated’ or ‘not indicated’ is necessarily agent relevant. At the core of CSMR is not that there is no maternal of fetal indication in absolute terms, but rather that (some) doctors do not share the same valuation of risk–benefit as viewed by the patient or that they do not regard the risk–benefit ratio favourable for the performance of caesarean section. As mentioned above, at the time when caesarean section was associated with high maternal mortality, Meigs articulated an opinion against caesarean section for any fetal indication. But the fact that the safety profile of caesarean section has changed is apparent to all, resulting in a shift of focus to quality of life considerations. Common reasons for women to request CSMR include the desire to avoid labour pain and stress, the wish to avoid uncertainty, fear of emergency interventions and the need to maintain a level of control, fear of forceps, concerns about fetal well being including the wish to avoid trauma or fetal distress in labour, as well as factors related to vaginal prolapse and urinary incontinence. Whilst literature may be able to provide a numerical estimate of these occurrences, it remains impossible to understand the value each individual woman places on them without directly seeking her view.

Amongst the considerations commonly debated are those relevant to vaginal function and continence. It is clear, including to ordinary people, that vaginal birth affects vaginal and perineal anatomy and that it results in ‘physiological’ perineal tears. Routine perineal incision or episiotomy has been abandoned in most obstetric practice, but the notion of a ‘cut’ or a ‘tear’ is recognised in lay language. Yet, fear of perineal trauma is cited as a good example of issues that ‘scare and undermine’ women’s ability to successfully undergo a normal process [20]. Various extrapolations and interpretations of statistics are often produced in this area [32–34]. Rortveit et al. studied the prevalence of urinary incontinence in women younger than 65 years [35, 36]. They reported that the adjusted odds ratio for any incontinence associated with vaginal deliveries as compared with caesarean sections was 1.7 (95% confidence interval, 1.3–2.1), and the adjusted odds ratio for moderate or severe incontinence was 2.2 (95% confidence interval, 1.5–3.1). Only stress incontinence (adjusted odds ratio, 2.4; 95% confidence interval, 1.7–3.2) was associated with the mode of delivery. Still, they concluded by emphasising their viewpoint that: ‘these findings should not be used to justify an increase in the use of caesarean sections’. It is interesting to note that those who oppose CSMR refer to patients’ ‘fear’ rather than their wish to avoid ‘risk’ of a particular complication. This helps foster the impression of a contrast with a more detached or rational medical view that is expressed using the language of ‘fact’ and ‘risk’.

Media interest in this topic remains high. On 12th April 2012, Reuters carried a news article reporting that: ‘Women who have given birth vaginally are more likely to develop incontinence decades later than moms who delivered their babies via cesarean section, according to a new study from Sweden’. This was accompanied by a comment from a practicing urogynaecology specialist stating that: ‘Anybody who has ever witnessed a vaginal delivery realizes the baby’s head is quite large and the muscles that it passes through are not that large. And any time you stretch a muscle there’s the potential for damage’ [37]. The study subject to this press interest ([38]) reported that two decades after one birth, vaginal delivery was associated with a 67% increased risk of urinary incontinence, and that urinary incontinence for more than 10 years increased by 275% for vaginal delivery compared with caesarean section [38]. The authors calculated that based on their data, it is necessary to perform eight or nine caesarean sections to avoid one case of urinary incontinence. They also found no difference in the incidence of incontinence between those who had an elective or an emergency caesarean section. This suggested that incontinence arises following the passage of the fetal head through the birth canal. Other studies have also linked caesarean sections to a reduced risk of pelvic floor disorders [39]. Perhaps not surprisingly, the issue remains hotly debated. But whilst debates are likely to continue, it is important to consider that even if an exact risk figure, or the estimate of the number needed to treat or to harm were to be agreed, this cannot determine what ought to be done at the level of the individual. Space does not allow an extensive discussion about each of the factors that are considered in literature or the media, but it is important to point out how often weak or inconclusive scientific content provides the context for sensational media reporting.

As discussed above, medical practice has shifted from the very restrictive early start to the stage where between one fifth and one third of all deliveries are conducted by caesarean section. The question must be asked as to why the insistence against accommodating maternal expressed wishes. A proposed answer may be that the professional view is a reflection of progress brought about through advancements in safety and that ‘medical indication’ is a reflection of where ‘evidence’ indicates a right balance which allows doctors to exercise their duty of beneficence and non-maleficence. This may seem plausible, except that assessments of benefit or harm are value judgements and, as such, are agency relevant. It has long been argued that doctors’ training does not qualify them to become arbiters of best interest. Indeed, as Veatch points out, it is difficult to argue that a physician who is expert in only one component of well-being is able to determine what constitutes the good for another person or to propose a plan to which individual patients would offer mere consent [40, 41]. A patient’s best interest is not an objective reality that could be elicited by a doctor based on the outcomes of clinical experiments performed on people with similar conditions, or based on the doctor’s own evaluation of whether a particular outcome, complication or risk is preferable to another. Irrespective of the theory of good adopted, it appears that the only way of knowing what is good for a patient is to ask her individually. The idea that a clinician can determine what is a ‘medically indicated’ intervention or what is in the patient’s ‘best interest’ must reside either in paternalism or reflect a misunderstanding of what a clinician can do [40]. The other critical factor in the debate concerns the place and valuation placed on autonomous choice and patient expressed preferences.

Paternalism is perhaps one of the more common criticisms levelled at the medical profession [42] and is one that is difficult to defend. The imposition of benefit is necessarily paternalistic, and this remains true irrespective of the nature or magnitude of benefit. It is argued that paternalism is wrong because it violates autonomy, it is a violation of one’s perception of oneself, it is a hindrance to achieving self-determined objectives, or because it reflects lack of recognition of others as capable of independent choice. Berlin puts it as follows:

‘Paternalism is despotic, not because it is more oppressive than naked, brutal, unenlightened tyranny, nor merely because it ignores the transcendental reason embodied in me, but because it is an insult to my conception of myself as a human being, determined to make my own life in accordance with my own (not necessarily rational or benevolent) purposes…’ [43]

Autonomy is inextricably linked to the Western tradition of liberalism, and is given central status in Kantian moral philosophy and in Mill’s utilitarian liberalism [25, 50]. The principle of respect for autonomy requires that the views of those who are capable of deliberation about their personal goals be sought and respected.

The ascent of autonomy in medical ethics is relatively recent. Schneewind traced this to the end of the eighteenth century when there was a shift from the conception of morality as obedience to a conception where individuals were seen as equally able to live together in a morality of self-governance [44]. The emerging view was that all individuals are, in principle, equally able to recognise for themselves what morality calls for and to act accordingly. This conception, which was not confined to the clinical interface, came to challenge the earlier view that most people are not able to see what morality requires or to understand the reason for moral dictates. The older conceptions that gave rise to the need for higher authorities from which ordinary (or most) people obtain guidance or instruction linked to threats of punishment or promises of reward have thus been superseded.

Beauchamp and Childress noted that although respect for the autonomous choices of a person runs as deep in today’s common morality as any principle, there is little agreement about its nature, scope or strength [45]. They argued that autonomy should not be excessively individualistic, excessively focused on reason or unduly legalistic. They proposed that autonomy should allow for the social nature of individuals including their emotions, the impact of their choices on others, and that it should not be a mere front for the exercise of legal rights. Beauchamp and Childress also argued that at a minimum, respect for autonomy acknowledges the person’s right to hold views, to make choices, and to take actions based on personal values and beliefs, and that it also involves or requires from others respectful actions that go beyond non-interference in others’ personal affairs [45]. They argue that respect for autonomy entails acknowledging decision-making rights and enabling persons to act autonomously. The emphasis on autonomy within normative ethics generates a number of challenges to practising clinicians. A conflict may arise between the doctor’s view of their role, their desire to respect autonomous choice and their other ethical duties such as beneficence and non-maleficence. It ought also to be recognised that autonomy is necessarily restricted by the practical confines within which it could be exercised. It may be possible to resolve the potential difficulty posed by non-availability of willing care providers, but cost differentials and other practical relevant factors are grounded in the real world. Questions of distributive justice can feature prominently in debates about provision in public or other insurance-based health care systems because individual demand is not usually seen as sufficient grounds for care provision, primarily because of the likely burden on others. Whether an intervention is seen as medically justified, a matter of choice or as a resource-based determination will have a bearing on provision in privately funded services. Doctors have traditionally endeavoured to maximise patient benefit as entailed within the Hippocratic tradition and have thus been hesitant to positioning themselves as arbiters in decisions that are primarily concerned with resource. This does not imply that resource implications do not factor into doctors’ decision-making, arguably these ought to, but if cost or other practical considerations were the reason to limit or deny autonomous choice, this ought to be made explicit.

Current emphasis on autonomy may underpin those practices where patients are simply given information or a range of options and then left to choose. Arguably, this does not provide a convincing paradigm for the delivery of an obligation to benefit or to avoid harm and it would strain credibility to label a choice for caesarean section within such a construct as CSMR. Examples of this practice may have prompted criticism such as that by Hall and Schneider who argued that ethicists have moved towards what could be called mandatory autonomy or that patients should make their own decisions whether they want it or not or that the emphasis has shifted from what patients do want to what patents should want [46]. Empirical evidence is also advanced to support the idea that at least some patients do not want such ‘unwanted’ autonomy. But presenting examples where doctors are unwilling or unable to exercise their duties as a ‘triumph of autonomy’ can potentially mask the realisation that current practice readily accommodates patient choice only if that falls within the range of options predefined or delimited by the doctor. In today’s practice, patients are seen to be free to accept or to refuse any of the options offered but barriers emerge against expressed preferences if these fall outside the orthodoxy.

It is though critically important that the medical encounter is not reduced to an interaction through which doctors simply provide learnt technical skill in response to determination by patients. Such would constitute a fundamental departure from the duties entailed within the Hippocratic Oath and subsequent medical codes of practice. There is a substantial risk that it would be detrimental to patients if their care were to be dictated by them, not because they are not the best arbiters of their needs, but because they usually lack the depth of knowledge or expertise that enables them to fully appreciate the implications of the various modalities of treatment. It is the need for such experience that drives patients to seek medical care. It would also be important to ensure that patients are not under undue influence or misconceptions when expressing their choices. Relevant to this are difficulties and challenges linked to providing non-directive counselling. It could be seen how preferences or biases held by clinicians or others can operate, covertly or overtly, to affect patients’ choice or expressed requests. This is an area where safeguards are needed.