NEGLIGENCE — Medical negligence — Peer professional opinion — Civil Liability Act 2002 (NSW) s 5O — Whether necessary that at the time of the alleged negligence there was an established practice which was widely accepted as competent medical practice — McKenna v Hunter & New England Local Health District  NSWCA 476; (2013) Aust Torts Rep 82–158 considered.
 Australian authority, as explained by Gleeson CJ in Rosenberg v Percival,15 adopted the dissenting view of Lord Scarman by which “the relevance of professional practice and opinion was not denied; what was denied was its conclusiveness.” The Australian approach was not confined to cases of warnings and medical advice; it extended to diagnosis and treatment.16
 However, the language of s 5O differs from Bolam, which referred to a practice accepted as “proper” by a “responsible body of medical men”. Lord Scarman also referred to “responsible and competent professional opinion”. The differences may not have practical significance.
 In Dobler, Giles JA stated :17
Section 5O may end up operating so as to determine the defendant’s standard of care, but the standard of care will be that determined by the court with guidance from evidence of acceptable professional practice unless it is established (in practice, by the defendant) that the defendant acted according to professional practice widely accepted by (rational) peer professional opinion.
Ipp JA and I agreed. On reflection, this passage may be open to misunderstanding. It is true that s 5O will not be engaged unless there is evidence of a widely accepted professional practice supporting the defendant’s conduct, but where there is such evidence, unless it can be rejected by the trial judge, it will fix the relevant standard; there cannot be two legally supportable standards operating in the one case.
 There remains a question as to the extent to which the trial judge has a discretion to reject evidence of a widely accepted professional practice. The judge may dismiss “peer professional opinion” as “irrational”, in accordance with subs 5O(2). By contrast, the older cases, while tending to use transferred epithets, refer to opinions of responsible and competent practitioners in the field. It may be that the reference in s 5O to “peer professional opinion” is to be so understood. But then it is difficult to understand how an irrational opinion could qualify as a relevant peer opinion.
 More importantly, the provision raises the possibility of a negative inference, namely that the court may not reject an opinion even though satisfied that it is unreasonable (though not irrational), or otherwise not one the court would itself adopt. Such a general negative inference should not be inferred; there will be other questions which will arise.
 First, there will be a question as to whether the evidence of one or two experts can satisfactorily establish opinions which are “widely accepted” in circumstances where such a view is contradicted by other evidence. No doubt evidence of “general professional opinion”, in addition to the personal opinion of the expert, is admissible in such circumstances.18
 Secondly, it will be a matter for the court to assess the significance of particular evidence. Evidence may be at a greater or lesser level of generality. At a high level of generality it may readily be accepted that an opinion is widely held amongst peers of the practitioner. However, the standard so identified may not assist greatly in resolving the particular case. On the other hand, the more particular the opinion, based on the specific circumstances of the case, the more difficult it may be to establish an opinion which can be described as “widely accepted” among fellow practitioners. Accordingly, whether or not evidence of medical opinion is properly described as conclusive in a particular case will depend upon a range of factors and not merely the fact that it can properly be described as not irrational.
 Subsections (3) and (4) do little more than explain what might otherwise readily have been inferred, namely that for an opinion to be “widely accepted” it need not be “universally accepted” and, if not universally accepted, there must be more than one opinion and may be more than one which is widely accepted. If the court is satisfied that the defendant’s conduct is considered competent according to widely accepted peer professional opinion, that opinion provides a standard against which to assess the claim.
 At trial, both practitioners raised s 5O as a defence. Counsel for the plaintiff treated the reasoning of this Court in McKenna v Hunter & New England Local Health District19 as an effective answer to the defence. In McKenna Macfarlan JA stated that:20
To establish a defence under s 5O a medical practitioner needs to demonstrate, first, that what he or she did conformed with a practice that was in existence at the time the medical service was provided and, secondly, to establish that that practice was widely, although not necessarily universally, accepted by peer professional opinion as competent professional practice.
On the appeal, counsel for Dr Sparks submitted that McKenna was wrong in holding that it was necessary to demonstrate the existence of “a practice” extant at the time of the conduct in question.
 Although it is uncertain how this reasoning in McKenna will operate in particular cases, there is a risk in reformulating the statutory language. To speak of “a practice” adopted by a group of professional persons suggests a regular course of conduct adopted in particular circumstances. By contrast, the phrase “competent professional practice” is apt to cover the whole gamut of professional services provided by the practitioner, whether or not the particular circumstances have arisen sufficiently often to result in an established practice. For example, although opinions may differ as to the conclusion to be drawn, there is no grammatical or semantic difficulty in describing an argument run by counsel in a novel case as demonstrating competent or incompetent professional practice. The same judgment may be offered about the failure of counsel to call a defendant in a criminal trial, where no settled practice exists. Where an acquittal depends on establishing an affirmative defence and there is no other evidence to support the defence, it may be described as incompetent professional practice not to call the defendant who could have given such evidence. Where an opinion is given and challenged, it will be supported (or attacked) not by reference to some established practice, but by reference to how an assessment of the circumstances (which may be unique) would be undertaken by a knowledgeable and experienced practitioner.
 There are other reasons for thinking that the reference to “competent professional practice” does not require evidence of “a practice”. First, it is the “manner” in which the defendant acted which must be the focus of the opinion. Secondly, if it were necessary to establish a practice, one might expect subs (3) to refer to “opinions … concerning that practice”, rather than “opinions … concerning a matter”.
 To take an example closer to the present case (but still hypothetical) an anaesthetist might allow an operation to proceed on the basis that two indicators remained within acceptable limits but a third indicator did not. An expert might express an opinion that such conduct was not competent practice, not because he or she had experienced the same circumstance in the past, or had read about it in a textbook, but because basic principles of human physiology led to that conclusion.
 Accordingly, although the language used in McKenna may well sufficiently describe many circumstances in which s 5O is invoked, I would not understand it as a general proposition as to the constraints imposed by s 5O(1).
 If that understanding is too restrictive and it is necessary to go further, I would not follow McKenna. The decision in McKenna having been overturned in the High Court,21 the reasoning of the majority in this Court is no longer binding. As Kirby J noted in Garcia v National Australia Bank:22
It is fundamental to the ascertainment of the binding rule of a judicial decision that it should be derived from (1) the reasons of the judges agreeing in the order disposing of the proceedings; (2) upon a matter in issue in the proceedings; (3) upon which a decision is necessary to arrive at that order. Thus, the opinions of judges in dissent are disregarded for this purpose, however valuable they may otherwise be.
 When the decision of an intermediate court of appeal is overturned by the High Court, the reasoning of the majority is no longer dispositive, even if the High Court does not directly reject it, as it did not with respect to the interpretation of s 5O in McKenna, having determined the appeal on a different point.
 There is some irony in the fact that the primary support for this conclusion is often sourced to a brief statement by Aickin J in dissent in Federal Cmr of Taxation v St Helens Farm (ACT) Pty Ltd23 stating:
I should add that there is no basis on which one point in the judgment of a primary court should be regarded as authoritative where the judgment is reversed on other grounds.
 Given that the reasons of the intermediate court in a case where the decision has been reversed are no longer dispositive, they are analogous to the reasoning of a dissenting judge. Allsop P explained in Holmes a Court v Papaconstuntinos24 that dissenting judgments “may contain valuable discussions of legal principle”, but that is “a different thing to being taken as an exposition of the common law to be applied.” The same proposition applies with respect to discussion of statutes.
 The position would, of course, be quite different if the High Court, despite reversing the decision in this Court in McKenna, had approved the reasoning with respect to s 5O. The reasoning would not then form part of the ratio, but it would clearly obtain the authority of dicta of the High Court. However, that did not happen in the present case. The High Court held that the plaintiff must fail on the basis that there was no duty of care owed to relatives of the deceased. The Court continued:25
Consideration of those other issues, about ss 5B, 5O, 43 and 43A of the [Civil Liability Act], should await a case in which it is necessary to examine them.
 While the reasoning of this Court in McKenna as to the scope and operation of s 5O is not to be disregarded, there is no obligation on this Court, as a matter of precedent, to follow it. For the reasons set out above, in my view it is too restrictive.26