S259/2009

IpsofactoJ.com: International Cases [2010] Part 8 Case 15 [HCA]


HIGH COURT OF AUSTRALIA

Coram

Tabet

(by her tutor, Ghassan Sheiban)

- vs -

Dr Gett

GUMMOW ACJ

BELL J

CRENNAN J

HAYNE J

HEYDON J

KIEFEL J

21 APRIL 2010


Judgment

Gummow ACJ

  1. The appellant suffers irreversible brain damage. She was born in 1984 and brought her action in the Supreme Court of New South Wales by her tutor, who is her uncle. The respondent, Dr Gett, was the second defendant in the action and at all material times was a registered medical practitioner practising as a paediatrician and a visiting medical officer at the Royal Alexandra Hospital for Children in Sydney. The events giving rise to the litigation took place when the appellant, then aged six years, was a patient at the hospital. Changes to the common law subsequently made by the Civil Liability Act 2002 (NSW) did not apply directly to this case.

  2. The action was brought in negligence alone and there was no claim in contract. The appellant pleaded her case as one in which breach by the respondent of his duty to manage her with due care and skill caused or contributed to cause her injury, loss and damage, or, in the alternative, led to "the loss of an opportunity to avoid injury, loss and damage." The appellant at trial failed on the first ground but succeeded on the second.

  3. There are two central issues. The first is whether in a claim arising from personal injury the law of negligence permits the bifurcation in this way of the nature of the actionable damage attributable to the same breach of duty, so that failure of the case on the first branch may be overcome by success on the second. In substance, the respondent contends that these are not true alternatives and that the law of negligence does not recognise as compensable damage the loss of opportunity in question here. The second issue is whether, in any event, the evidence sufficiently supported the favourable finding at trial on the claim for loss of opportunity.

    The course of the litigation

  4. On 11 January 1991 the appellant was admitted to the hospital and came under the care of the respondent. The appellant had recently suffered from chickenpox which had resolved but both before and after that illness she suffered from headaches, nausea and vomiting. The respondent made a provisional diagnosis that the appellant was suffering from chickenpox, meningitis or encephalitis.

  5. The significant events which followed, and the course of the 36-day trial before Studdert J (Tabet v Mansour [2007] NSWSC 36), were described as follows by the Court of Appeal (Allsop P, Beazley and Basten JJA: Gett v Tabet (2009) 254 ALR 504 at 506-507):

    On 14 January 1991, after suffering a seizure, and after a CT scan and EEG were performed [the appellant] was diagnosed as suffering from a brain tumour. She received treatment, including an operation to remove the tumour. She suffered irreversible brain damage, partly as a result of events on 14 January 1991, partly from the tumour (which had been growing for over 2 years), and partly from the operative procedure and other treatment (not said to be in any way negligently performed).

    The [appellant] brought proceedings against the [respondent] in negligence. The central allegation was that the CT scan that was undertaken on 14 January should have been performed earlier, either on 11 or 13 January, and that if it had, she would have had a better medical outcome. The plaintiff also brought proceedings against Dr Mansour, who had treated her in an earlier admission to hospital on 29-31 December 1990. The trial judge held that Dr Mansour was not negligent in his treatment of the [appellant] and there is no appeal from that decision.

    The trial judge held, however, that the [respondent] was negligent in failing to order a CT scan on 13 January 1991. His Honour found no earlier negligent act or omission, thereby concluding that the [respondent] acted reasonably in making his provisional diagnosis on 11 January that the [appellant] was suffering from chickenpox or varicella meningitis or encephalitis.

  6. The finding by Studdert J of negligence in failing to order a CT scan on 13 January was based upon an episode at 11 am on that day when nursing staff observed that the appellant's pupils were unequal and the right pupil was not reactive. However, his Honour was not persuaded on the balance of probabilities that the discovery of the tumour upon administration of a CT scan on 13 January would have led to the appellant being treated in such a way as would have avoided the seizure and deterioration in her condition on 14 January.

  7. The Court of Appeal continued its account of the trial as follows (ibid at 507):

    Having found that the [respondent] breached his duty of care, the trial judge did not conclude that this negligence caused or contributed to the seizure and deterioration which occurred on 14 January. Rather, his Honour found that the [appellant] lost a chance of a better medical outcome had the brain tumour been detected on 13 January 1991, as it would have been if the CT scan had been performed that day.

  8. The trial judge had introduced his holding with respect to the loss of a chance by stating that on the balance of probabilities he was satisfied that, had a CT scan been called for at 11 am on 13 January, it would have been performed urgently, the tumour would have been detected and treatment, probably by administration of steroids rather than by drainage, would have reduced intracranial pressure. The absence of treatment "deprived [the appellant] of the chance of a better outcome" [2007] NSWSC 36 at [306]. Further, the detection of the tumour on 13 January would have eliminated the time lost in carrying out the CT scan after the seizure on the next day and before urgent surgery was subsequently performed.

  9. There was then the question of remedy. The Court of Appeal described the outcome of the trial as follows ((2009) 254 ALR 504 at 507):

    His Honour assessed the [appellant's] damages referable to her entire brain damages in a total amount of $6,092,586. His Honour found that it was probable that the [appellant's] decline on 14 January contributed to her ultimate disabilities and assessed that contribution to be no greater than 25%, representing an assessment of $1,523,146. An attack by the [respondent] on this divisible apportionment was abandoned at the appeal. The trial judge assessed that the loss of a chance of a better outcome, that is avoiding the damage referable to the deterioration on 14 January (the 25%), was 40%. The damages to which the [appellant] was thus entitled for a 40% loss of a chance was $610,000. His Honour thus ordered verdict and judgment for the [appellant] in that sum.

  10. The trial judge emphasised that the "loss of a chance" branch of the appellant's case, upon which alone she had succeeded, had not been her primary claim. This had been for recovery for negligence resulting in her brain damage and the appellant's case had been that, even if this result were restricted to the harm suffered on 14 January, her overall disability was indivisible. It may be accepted that the medical evidence had been led and cross-examined with that primary claim uppermost in mind.

  11. The Court of Appeal set aside the judgment for the appellant and entered judgment for the respondent. The appellant in this Court seeks the restoration of the outcome at trial. The principal ground of appeal is that the Court of Appeal erred in holding that the causal effects of the clinical negligence of the respondent should be assessed on the balance of probabilities alone rather than, as at trial, "on the basis of loss of a chance of a better outcome." That is to say, the appellant disputes the adverse outcome for her in the Court of Appeal on what earlier in these reasons is identified as the first of the central issues.

  12. It should be noted that in the Court of Appeal,

    1. the respondent failed in his challenge to the finding that he had been negligent in failing, on 13 January, to consider other possible diagnoses and to order a CT scan on that day, and

    2. the appellant failed in her contention that she had suffered more than the loss of the opportunity of a better outcome and that the primary judge should have found, on the balance of probabilities, that the negligence of the respondent had caused the whole of the brain injury referable to her seizure and deterioration on 14 January, being 25 per cent of her overall disability after the operation.

    The state of authority

  13. In ruling in favour of the appellant on the "loss of a chance" branch of her case, the trial judge drew support from the decisions of the Victorian Court of Appeal in Gavalas v Singh (2001) 3 VR 404 and of the New South Wales Court of Appeal in Rufo v Hosking (2004) 61 NSWLR 678. But, as Callaway JA emphasised in the first case ((2001) 3 VR 404 at 409), the appeal had turned upon the assessment of damages and, as M W Campbell AJA explained in the other case ((2004) 61 NSWLR 678 at 693-694), the litigation there was conducted on the basis that if the facts supported a claim based on the loss of a chance then the action lay. Nevertheless, in the present case Studdert J regarded himself as bound by Rufo.

  14. Shortly after Rufo was decided by the New South Wales Court of Appeal, in Gregg v Scott [2005] 2 AC 176 the House of Lords (Lords Hoffmann and Phillips of Worth Matravers MR and Baroness Hale of Richmond; Lords Nicholls of Birkenhead and Hope of Craighead dissenting) affirmed the rejection by the Court of Appeal (Simon Brown and Mance LJJ; Latham LJ dissenting) ([2003] Lloyd's Rep Med 105) of the submission by the plaintiff that the trial judge should have awarded him damages on the footing that the reduced chances of successful treatment of the cancer he suffered should be recoverable as damages in negligence.

  15. In breach of his duty of care owed to his patient Mr Gregg, Dr Scott had failed forthwith to refer him for a biopsy investigation in November 1994, and the cancer was undetected until November 1995. There was a delay in the commencement of treatment, in which time the cancer had spread. But Mr Gregg had been in remission since 1998. The trial was in 2001 and there was no discernible recurrence of the disease as the litigation proceeded to the House of Lords. Expert evidence treated a "cure" as disease-free survival for 10 years. It was an agreed fact that had Mr Gregg been promptly diagnosed and treated his chance of disease-free survival for 10 years would have been 42 per cent, but at the trial in 2001 this had been reduced to only 25 per cent.

  16. Mr Gregg was thus a survivor against what might be called the statistical odds. His sole complaint was that the breach of duty by Dr Scott had reduced his prospects of a "cure", being disease-free survival until at least 2008. However, the chronology meant that the chance had not yet run its course. It thus remained unsettled whether Dr Scott's breach of duty had destroyed the chance of a "cure." It may have been a paradox that Mr Gregg's resilience made it more difficult for him to establish his case. Nevertheless, how, it might be asked, had the damage, the loss of the chance which was the gist of the action, yet been sustained?

  17. Moreover, there was a risk of over-compensation if Mr Gregg recovered damages upon his action tried in 2001 for the reduction in his prospects of survival by reason of the negligent failure in diagnosis in 1994: Cf Wardley Australia Ltd v Western Australia (1992) 175 CLR 514 at 527. As a general proposition, and in many fields of law, assessments of compensation or value are made by taking into account all matters known at the later date, when conjecture is no longer essential: HTW Valuers (Central Qld) Pty Ltd v Astonland Pty Ltd (2004) 217 CLR 640 at 658-659 [39].

  18. Against that background, it is, with respect, unsurprising that one of the majority, Lord Phillips, said that it would be unsatisfactory to award damages for the reduction of the chance of a cure when the long-term result of treatment is still uncertain ([2005] 2 AC 176 at 225[1]), thereby perhaps threatening the coherence of the common law: [2005] 2 AC 176 at 221[2]. A similar concern for coherence in the tort of negligence is apparent in the opinion of Baroness Hale. She asked how a personal injury law concerned with outcomes could live with an alternative of recovery for loss of a chance of an outcome: [2005] 2 AC 176 at 233.

  19. Counsel for the present appellant, seeking to diminish the significance for this case of the reasoning of the majority in Gregg v Scott, emphasised that, unlike Gregg v Scott, it does not present the conundrum of a chance or prospect of the plaintiff dying earlier than would otherwise be the case as the basis for an action brought while the plaintiff still lived. Here, it was said, the end result, the appellant's disabilities, had been reached before the action was commenced.

  20. Earlier, in Laferrière v Lawson [1991] 1 SCR 541 the Supreme Court of Canada, on appeal from the Quebec Court of Appeal ([1989] RJQ 27), had considered "loss of chance" in medical negligence. The reasons of the majority (La Forest J dissenting) were given by Gonthier J, who said of "the loss of chance analysis" that it added unnecessary and impermissible confusion to medical negligence cases because it "in fact hides a break in the causal link" : [1991] 1 SCR 541 at 591. However, in the British Columbia Court of Appeal it has been said by Southin JA that, in that Province, the relationship of patient and physician is essentially contractual. The patient has the right to performance of the contract on its terms and on that basis there might be recovery of damages representing the loss of a chance of less than 50 per cent of a better outcome[3]. But, as indicated above, there was no contractual claim in this case and no occasion to consider the approach taken by Southin JA: Cf Breen v Williams (1996) 186 CLR 71; Wong v The Commonwealth (2009) 236 CLR 573.

  21. Perhaps more immediately congenial to the appellant's case is the recent decision of the seven member Supreme Judicial Court of Massachusetts in Matsuyama v Birnbaum 890 NE 2d 819 (2008)[4]. By reason of the failure in diagnosis by the defendant in 1995, Mr Matsuyama's cancer, which then might still have been curable, had metastasised to an advanced inoperable phase resulting in his premature death: ibid at 826 (2008). In upholding the finding of the jury that the misdiagnosis was a "substantial contributing factor" to the death of Mr Matsuyama, Marshall CJ said (ibid at 823 (2008)):

    the loss of chance doctrine views a person's prospects for surviving a serious medical condition as something of value, even if the possibility of recovery was less than even prior to the physician's tortious conduct. Where a physician's negligence reduces or eliminates the patient's prospects for achieving a more favorable medical outcome, the physician has harmed the patient and is liable for damages. Permitting recovery for loss of chance is particularly appropriate in the area of medical negligence. Our decision today is limited to such claims.

    Her Honour went on to stress that if "loss of a chance" is to be recognised as actionable it is better understood as an injury recognised by the law of tort, than as a separate cause of action or as a surrogate for the necessary element of causation in a negligence claim: ibid at 832 (2008) (adopting Alexander v Scheid 726 NE 2d 272 at 279 (2000)). If recovery be sought for decrease in the patient's prospect of recovery, rather than the ultimate outcome, there has to be identification and valuation of that diminished prospect. With that I, with respect, agree. But that does not mean that issues of causation do not arise on such an analysis. It will be necessary to say more of this important consideration later in these reasons.

  22. However, as Kiefel J explains in her reasons, the form of the actual recovery in Matsuyama was in the controversial shape of "proportional damages" representing not the loss of a chance of survival but a percentage of a damages award on a statutory wrongful death claim by the executrix of Mr Matsuyama.

  23. In Matsuyama, Marshall CJ did emphasise two matters, both of which are uncontroversial and applicable to the present appeal. The first is the importance of determinations of fact based upon expert testimony rather than speculation based on insufficient evidence: ibid at 833-834 (2008). The second is the distinction between the injury or damage which is the gist of the action in negligence and the proper measure of damages: ibid at 838-839 (2008). Much of the difficulty derives from the multiple reference of the term "damage", which is used to identify that which the law accepts as sufficient injuria, and the measure of compensation represented by the sum for which judgment is entered[5].

  24. In that regard Lord Walker of Gestingthorpe has emphasised that while questions of assessment of damages may involve quantifying future or hypothetical chances, the common law has not accepted that the attribution of liability should be proportionate to the proof of causation[6]. Nevertheless, in some of the cases there has been a tendency to run together questions of attribution of liability and the measure of damages recoverable.

  25. In the present case, the Court of Appeal considered, and properly so, that it could only be for this Court "to reformulate the law of torts to permit recovery for physical injury not shown to be caused or contributed to by a negligent party, but which negligence has deprived the victim of the possibility (but not the probability) of a better outcome."

  26. Their Honours added ((2009) 254 ALR 504 at 586):

    Such an approach would not readily be limited to medical negligence cases, but would potentially revolutionise the law of recovery for personal injury. It would do so by reference to an assessment of increased risk of harm, verbally reformulated into loss of a chance or opportunity in order to equate it with the recognition in Sellars [v Adelaide Petroleum NL [(1994) 179 CLR 332] and like cases of the existence in commerce of a coherent notion of loss of a right or chance of financial benefit. No doubt the limits of the 'commercial' or financial opportunity or advantage dealt with in Sellars will be a matter of future debate: see the discussion in Gregg [v Scott [2005] UKHL 2; [2005] 2 AC 176 at 232] by Baroness Hale of Richmond]. In our view, its limits (unless expanded by the High Court) must fall short of a proposition which revolutionises the proof of causation of injury or [which redefines what is 'harm'] in personal injury cases.

  27. These reasons will seek to demonstrate that the reformulation of which the Court of Appeal spoke should not be made, and that the appeal to this Court must fail. However, this outcome will not require acceptance in absolute terms of a general proposition that destruction of the chance of obtaining a benefit or avoiding a harm can never be regarded as supplying that damage which is the gist of an action in negligence.

    The case for the appellant

  28. Studdert J had described the success of the appellant as the entitlement "to be compensated for the loss of a chance of a better outcome had the breach of duty not occurred": [2007] NSWSC 36 at [379]. The identification of the chance lost as that of a "better outcome" is repeated elsewhere in the reasons. As noted by the Court of Appeal, the "better outcome" appears to have been avoidance of the brain damage referable to the deterioration on 14 January. This was assessed as a contribution of 25 per cent to the ultimate disabilities which the appellant suffers. The "chance" of avoiding that brain damage referable to the deterioration of 14 January was assessed by the trial judge as 40 per cent.

  29. In this Court counsel for the appellant submitted that at the heart of her case was the concept of the lessening of the gravity of the final result. As refined in the course of argument, the appellant's case is that the respondent's negligence deprived her of a chance, prospect or opportunity that had remained open only for a short period between 11 am on 13 January and her seizure and deterioration on 14 January. The chance, prospect or opportunity had been of avoiding so much of the eventual outcome, her disabled state, which as to 25 per cent was attributable to her seizure and deterioration on 14 January.

  30. The appellant sought to stigmatise the respondent's case as being that, because the likelihood of this better outcome was less than 50 per cent, it followed (a) that on the balance of probabilities the appellant would still have suffered as much as she did, and therefore (b) the chance, prospect or opportunity had no worth.

  31. However, if the likelihood of a better outcome had been found to be greater than 50 per cent then on the balance of probabilities the appellant would have succeeded, not failed, on the main branch of her case in negligence. The question of principle thus becomes whether the law permits recovery in negligence on proof to the balance of probabilities of the presence of something else, namely a chance, opportunity or prospect of an outcome the eventuation of which, however, was less than probable.

    The case for the respondent

  32. The respondent submitted that even if (which he disputed) the appellant had correctly formulated the applicable legal principles, the evidence had provided an insufficient basis for a favourable outcome based on anything more than speculation. The respondent also challenged the indeterminacy of the terms used by the trial judge, "better outcome" and "chance." The respondent submitted that the "chance" found was that steroids if administered, or a drain if inserted earlier, would have worked to lessen or avert brain damage, but that, in the way the evidence was led at the trial, these mere possibilities were not tied to evidence sufficient to found any assessment of the potential effectiveness of that chance. Indeed, at the trial, counsel for the respondent had submitted that there was no expert evidence as to the value of the lost chance or sufficiently identifying the actual harm suffered on 14 January.

  33. For the reasons which follow, the case presented by the respondent should be preferred to that presented by the appellant. This is so both with respect to the applicable principles, and with respect to what the respondent submits in any event to have been the weaknesses in the evidence.

  34. Before turning to matters of deep principle, something more should be said respecting the reasons of the trial judge.

    The reasons of the trial judge

  35. The trial judge began with the proposition, which the appellant properly accepted in submissions to this Court, that the existence of the chance of a better outcome had to be proved on the balance of probabilities. His Honour continued, ibid at [377], [378], [381], [382], [429]:

    I am satisfied on the probabilities as to the following:

    (i)

    that the failure to relieve the [appellant's] intracranial pressure during the twenty-four hour period prior to [her] decline on 14 January 1991 was causative of brain damage occurring at and following the time of the [appellant's] observed decline on 14 January;

    (ii)

    that the [appellant] lost the chance of that relief and of avoiding or minimising that damage by reason of the breach of duty of the [respondent];

    (iii)

    that the chance was not so low as to be speculative, but was a good chance, although less than a fifty percent chance.

    .... Whilst I do not have the benefit of any expert opinion as to the value of the lost chance in percentage terms, I have decided that the loss of the chance had the breach of duty not occurred is to be measured at forty percent. In so concluding, I have regard to the following:

    (a)

    the probability that a CT scan if sought would have been performed urgently, on 13 January, revealing the presence of the medulloblastoma;

    (b)

    the probability that the detection of the medulloblastoma would have been followed immediately on 13 January by the placement of a drain or the prescription of steroids. ... I consider it more likely that steroids would have been prescribed rather than the placement of a drain;

    (c)

    whilst I consider that the placement of a drain would have proved more effective to relieve pressure, the probability is that the steroids would have had some beneficial effect, and would have reduced tumour related swelling;

    (d)

    the carrying out of the CT scan on 13 January would have avoided the time occupied in performing the CT scan and the EEG on 14 January. Should it still have proved necessary to insert the drain on 14 January, notwithstanding the prescription of steroids the day before, this procedure could have been carried out approximately two hours earlier than it was. ...

    (e)

    If, notwithstanding the prescription of steroids on 13 January, the decline had still occurred on 14 January, the elimination of the delay for the CT and the EEG on the later date and the earlier insertion of the drain would have increased the chance of a better outcome. ...

    After close attention to the matter and whilst acknowledging the difficulties of the task, I have decided that I should proceed with the assessment [of the value of the lost chance].

    The [appellant] is entitled to damages referable to the loss of a chance of a better outcome in relation to the harm suffered on 14 January 1991 only. There were altogether four contributors to the totality of the brain damage from which the [appellant] presently suffers:

    (i)

    the medulloblastoma with its seeding, and the hydrocephalus;

    (ii)

    the damage that occurred on 14 January 1991;

    (iii)

    the surgery on 16 January 1991;

    (iv)

    the subsequent radiotherapy treatment.

    ....

    Having considered all the medical evidence, I think it probable that the event of 14 January made some contribution to the [appellant's] ultimate disabilities, particularly her cognitive loss and her ataxia, her loss of balance and her coordination impairment. However, I find on the probabilities that the contribution made by the event of 14 January 1991 to the above specified disabilities and to her disabilities generally was significantly less than the combined contribution of the remaining contributors. It is impossible to be precise about the matter, as reflection on the medical evidence reveals, but I find on the probabilities that the contribution of the event of 14 January 1991 to the aggregate brain damage and resulting disabilities with which the [appellant] has presented to this Court is no greater than twenty-five percent.

    [emphasis added]

  36. The Court of Appeal ((2009) 254 ALR 504 at 555, however, held that if, contrary to its view, a loss of chance analysis were "legitimate" it would consider that the appellant lost, at most, a 15 per cent chance, not a 40 per cent chance, of avoiding the overall 25 per cent of the brain damage.

    The evidence

  37. The trial judge plainly had appreciated the difficulty in deriving from the evidence the conclusions he reached both as to the 25 per cent contribution to the appellant's disabilities and as to the 40 per cent chance of avoiding the deterioration on 14 January. This situation may be contrasted, for example, with that disclosed in Matsuyama 890 NE 2d 819 at 824-828 (2008), where there was before the jury extensive evidence by expert witnesses to support the opinion that the development of gastric cancer was classified into four distinct stages with each carrying a diminished chance of survival, measured by five disease-free years after treatment. It was the development of medical science to the point that, at least for some conditions, expert evidence could replace speculation that, in the view of the Massachusetts court, made it appropriate to recognise loss of chance as a form of injury, ibid at 834 (2008).

  38. No doubt the present case arose in very particular circumstances making it difficult to find the appropriate comparator or counter-factual. Usually this will require proof of what would have been the plaintiff's position in the absence of the breach of duty by the defendant. The difficulty in the present case arises from the substitution, for which the appellant contends, of loss of the chance of a better outcome for proof of physical injury, as the gist of the cause of action in negligence.

  39. The cases dealing with the assessment of the measure of damages, whether in contract or tort, are replete with exhortations that precision may not be possible and the trial judge or jury must do the best it can. The treatment in Malec v J C Hutton Pty Ltd (1990) 169 CLR 638 of the assessment of damages for future or potential events that allegedly would have occurred, but cannot now occur, or that allegedly might now occur, is an example. But in that case the claim giving rise to the assessment had been for physical injury, the contraction of a disease as a result of the negligence of the defendant. The imprecision allowed in the assessment of damages in such cases does not necessarily or logically apply where a claim for physical injury fails but is said to be saved by transmutation of the damage alleged into the loss of a chance of a better outcome.

  40. With that in mind, something should be said respecting McGhee v National Coal Board [1973] 1 WLR 1; [1972] 3 All ER 1008. That decision of the House of Lords on appeal from Scotland may be read as deciding, on orthodox grounds, that the negligence of the defendants had materially contributed to the personal injury of the pursuer. That characterisation later was disputed by the House of Lords itself in Fairchild v Glenhaven Funeral Services Ltd [2003] 1 AC 32, but it is unnecessary here to enter upon that debate. What is presently significant is that in the interim, when giving that orthodox reading of McGhee, Lord Bridge of Harwich in Wilsher v Essex Area Health Authority [1988] AC 1074 at 1090 had said of the speeches in McGhee that their conclusion manifested a "robust and pragmatic approach" to the drawing of a legitimate inference from "the undisputed primary facts of the case."

  41. This unremarkable use of language in the context of physical injury cases was then translated in Rufo into something more. In that case, the New South Wales Court of Appeal ([2004] NSWCA 391; (2004) 61 NSWLR 678 at 694 per M W Campbell AJA), after citing Lord Bridge's statement in Wilsher, concluded that if "a robust and pragmatic approach" were adopted to the primary facts of that case, then it was more probable than not that the chance of a better medical outcome was lost, with that chance being more than speculative or remote.

  42. In the present case, the trial judge proceeded in his assessment of the evaluation of the lost chance mindful of what he saw as the invitation in Rufo to take "a robust and pragmatic approach." But, as the respondent emphasised, this had been advocated by Lord Bridge with respect to the drawing of inferences from undisputed primary facts, whilst here there were deficiencies in the evidence necessary to support a finding for the appellant on a critical matter.

  43. This critical matter concerned what would have been the efficacy of steroid treatment in the short period of opportunity between the episode at 11 am on 13 January and the seizure the next day. The evidence in re-examination of one of the expert witnesses, the neurosurgeon Mr Ian Johnston, was as follows:

    Q

    What is your view as to whether or not the use of steroids would have had a role in avoiding the incident of the acute decline on 14 January 1991 in this patient?

    A

    Well, this would be absolutely a guess; I mean, it's entirely speculative. I don't think they would have been sufficiently effective under those circumstances of this particular patient, Reema, to do that, but I don't know.

    Q

    What was it about her condition that leads you to express that opinion?

    A

    Well it was primarily the pressure was primarily due hydrocephalus [sic], but, you know, tumour-related swelling steroids would have improved [sic]. So it an issue [sic] of which was the more important and by how much. That is very speculative, I have to say. I mean, you could certainly make an argument that they would have improved the situation and that they may have prevented the episode, but, as I say, nobody could answer that with any certainty.

  44. Mr Klug's evidence in cross-examination was that, while in non-acute conditions the administration of steroids was very effective, faced with the situation on 13 January he would have had to have made a very careful assessment of the condition of the appellant and may have used a combination of steroids and a ventricular drain if there had been a risk of serious deterioration. But Mr Klug was not asked for an opinion as to the efficacy of that treatment in the period before the time of the seizure on the next day.

  45. This evidence provided a basis for no more than speculation as to the loss of a chance of a better outcome whether assessed at 40 per cent or (as the Court of Appeal indicated) 15 per cent. For that reason the appeal to this Court should fail.

    The issue of principle

  46. Further, and as an additional ground of decision, in personal injury cases the law of negligence as understood in the common law of Australia does not entertain an action for recovery when the damage, for which compensation is awarded consequent upon breach of duty, is characterised as the loss of a chance of a better outcome of the character found by the trial judge in this case.

  47. It should be said immediately that the principles dealing with recovery of damages for breach of contract offer no appropriate analogy. The action for breach of contract lies upon the occurrence of breach, but that in negligence lies only if and when damage is sustained. This has significance for the application of limitation statutes. But it has the further and relevant importance identified by Brennan J in Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 at 359. This is that in a negligence action, unlike an action in contract, the existence and causation of compensable loss cannot be established by reference to breach of an antecedent promise to afford an opportunity.

  48. In a contract case the plaintiff should be entitled at least to nominal damages for loss of the promised opportunity. The jury in Chaplin v Hicks [1911] 2 KB 786 assessed at £100 (at the time a not inconsiderable sum) the damages for the breach found of the contractual obligation to take reasonable means to give the plaintiff an opportunity of presenting herself for selection by the defendant in a competition with 12 prizes of three-year theatrical engagements. The defendant, later Sir Seymour Hicks, was a well-known actor and theatrical manager in Edwardian London, who had built the Aldwych and Globe theatres, presented successful musical comedies, and discovered new talent, including that of the young P G Wodehouse as a lyricist[7]. With these matters in mind, it is readily seen that the plaintiff lost a chance of real value. The unsuccessful submission to the Court of Appeal by McCardie for the defendant ([1911] 2 KB 786 at 788-789) was that the only remedy was nominal damages, because substantial damages were so contingent as to be incapable of assessment.

  49. Chaplin v Hicks [1911] 2 KB 786 is authority for the proposition that if a plaintiff, by the breach of contract by the defendant, has been deprived of something which has a monetary value, there is to be an assessment of damages notwithstanding difficulty in calculation or impossibility of making an assessment with certainty. This Court, speaking in McRae v Commonwealth Disposals Commission (1951) 84 CLR 377 (at 411-412 per Dixon and Fullagar JJ) of Chaplin v Hicks, said that the broken promise in effect had been to give the plaintiff a chance and that she would have had a real chance of winning a prize, and thus that it was proper enough to say that the chance was worth something.

  50. But these considerations do not appear in the frame of reference for the present case. As Brennan J indicated in Sellars (1994) 179 CLR 332 at 359, in an action in tort where damage is the gist of the action, the issue which precedes any assessment of damages recoverable is whether a lost opportunity, as a matter of law, answers the description of "loss or damage" which is then compensable.

  51. In D'Orta-Ekenaike v Victoria Legal Aid (2005) 223 CLR 1 at 37 [101], McHugh J said:

    Reasonable foreseeability of physical harm is generally enough to impose a duty of care on a person who knows or ought reasonably foresee that physical harm is a likely result of his or her conduct. Liability will arise when the duty is breached and where there is a causal relationship between the breach and the harm.

  52. Further, harm to the interests of the plaintiff which is not sustained by injury to person or property, in the ordinary sense of those terms, nevertheless may qualify in at least some cases as the compensable damage consequent upon a breach of a duty of care as understood in the tort of negligence. The decisions in Hill v Van Erp (1997) 188 CLR 159 and Perre v Apand Pty Ltd (1999) 198 CLR 180 respecting recovery for "economic loss" are well-known examples.

  53. Where the act or omission complained of does not amount to interference with or impairment of an existing right, some care is needed in identifying the interest said to have been harmed by the defendant and said to be sufficient to attract the protection of the law in this field. The point was made by McPherson JA in Christopher v The Motor Vessel "Fiji Gas" (1993) Aust Torts Reports ¶81-202 at 61,967. That process of identification requires a sense of the existing and inherent principles of the law: Smith v Jenkins (1970) 119 CLR 397 at 418; Cattanach v Melchior (2003) 215 CLR 1 at 30-31 [64]-[65]. One of those principles favours the development of the common law, and in particular the tort of negligence, in a coherent fashion: Sullivan v Moody (2001) 207 CLR 562 at 579-581 [50]-[55]; Harriton v Stephens (2006) 226 CLR 52 at 123 [242]-[243].

  54. In the present case, with reference to what had been said in Sellars (1994) 179 CLR 332 when dealing with an action to recover "loss or damage" under s 82 of the Trade Practices Act 1974 (Cth) for contravention of s 52 of that statute, the Court of Appeal referred to the existence in commerce of a coherent notion of loss of a right of a chance of financial benefit. In that regard, the statement of principle by Brennan J in Sellars: ibid at 364 is significant:

    As a matter of common experience, opportunities to acquire commercial benefits are frequently valuable in themselves, not only when they will probably fructify in a financial return but also when they offer a substantial prospect of a financial return. The volatility of the market for speculative shares testifies to both the valuable character of commercial opportunities and the difficulty of assessing the value of opportunities which are subject to serious contingencies. Provided an opportunity offers a substantial, and not merely speculative, prospect of acquiring a benefit that the plaintiff sought to acquire or of avoiding a detriment that the plaintiff sought to avoid, the opportunity can be held to be valuable. And, if an opportunity is valuable, the loss of that opportunity is truly 'loss' or 'damage' for the purposes of s 82(1) of the Act and for the purposes of the law of torts.

  55. More generally, in Naxakis v Western General Hospital (1999) 197 CLR 269 at 278 [29] Gaudron J observed that while, "where no other loss is involved", there was no reason in principle why loss of a chance or commercial opportunity should not constitute damage for the purposes of tort law, different considerations apply where the risk has eventuated and there has been physical injury.

  56. Her Honour continued (ibid at 279 [32]):

    The notion that, in cases of failure to diagnose or treat an existing condition, the loss suffered by the plaintiff is the loss of chance, rather than the injury or physical disability that eventuates, is essentially different from the approach that is traditionally adopted. On the traditional approach, the plaintiff must establish on the balance of probabilities that the failure caused the injury or disability suffered, whereas the lost chance approach predicates that he or she must establish only that it resulted in the loss of a chance that was of some value.

  57. Several considerations thus are presented. One may be seen from the statement by Professor David Fischer made upon consideration in 2001 of decisions in the United States, Australia and other common law jurisdictions[8]:

    A major rationale for loss of a chance where plaintiff cannot prove traditional damage is that the chance of obtaining a benefit or avoiding a harm has value in itself that is entitled to legal protection. Thus, destruction of this chance ought to be regarded as damage giving rise to an actionable tort. Characterizing the damage as the loss of a chance of avoiding harm (or gaining a benefit) relieves the plaintiff of the burden of proving that the harm itself (or lost benefit itself) occurred. At the same time, the characterization preserves the requirement that plaintiff prove [damage] by the usual standard of proof. Note, however, that under the 'chance has value' characterization, it is often easier to prove actionable damage. It is usually easier to prove that defendant created a risk of harm (or a risk of loss of benefit) than to prove that defendant caused the harm itself (or benefit itself).

    [footnotes omitted]

  58. But why should the law favour the weakening of the requirement for proving causation such that, in the situation posited by Gaudron J (which is found in the present litigation), the plaintiff should have the benefit and the defendant the detriment of an easier proof of actionable damage for a negligence action?

  59. It may be said that the "all or nothing" outcome on the balance of probabilities leads to "rough justice." But the traditional approach in personal injury cases represents the striking by the law of a balance between the competing interests of the parties, and the substitution of the loss of a chance as the actionable damage represents a shift in that balance towards claimants. Again, there may be a view that, especially with respect to medical treatment, the substitution assists in the maintenance of standards where there is a less than even chance of a cure. This was a consideration which Baroness Hale adverted to in Gregg v Scott [2005] 2 AC 176 at 231. But any such potential benefit to the public weal has to be weighed against, for example, the prospect of "defensive medicine" with emphasis upon costly testing procedures in preference to a sequential deductive approach to diagnosis and treatment.

  60. In Gregg v Scott, ibid at 231-232, Baroness Hale went on:

    But of course doctors and other health care professionals are not solely, or even mainly, motivated by the fear of adverse legal consequences. They are motivated by their natural desire and their professional duty to do their best for their patients. Tort law is not criminal law. The criminal law is there to punish and deter those who do not behave as they should. Tort law is there to compensate those who have been wronged. Some wrongs are actionable whether or not the claimant has been damaged. But damage is the gist of negligence. So it can never be enough to show that the defendant has been negligent. The question is still whether his negligence has caused actionable damage .... In this case we are back to square one: what is actionable damage?

  61. In that situation, it should be remembered that the duty of care and its breach are assumed. The determination of the existence and content of a duty of care is not assisted by looking first to the harm sustained by the plaintiff and then reasoning, as it were, retrospectively: Vairy v Wyong Shire Council (2005) 223 CLR 422 (at 443 [60]-[61], 461-462 [126]-[128]). Nor is it appropriate to reason that, duty and breach being established, the plaintiff who on the balance of probabilities cannot establish actionable damage nevertheless must have a remedy.

  62. Finally, there is the consideration which weighed with Gaudron J in Naxakis (1999) 197 CLR 269 at 280-281 [36], Gonthier J in Laferrière [1991] 1 SCR 541 at 605 and Lord Hoffmann in Gregg v Scott [2005] 2 AC 176 at 196. Where, as in the present case, and unlike in Gregg v Scott itself, the relevant risk of a bad outcome in a pre-existing but undiagnosed or untreated condition has eventuated before the institution of the litigation, the factors bound up in the earlier chance have played themselves out. What is in issue is past events, preceding in this case disabilities from which the appellant suffers. The cause of the disabilities, on the evidence, may be uncertain. But the difficulty which this presents is not overcome by removing the analysis of the facts and law to the more abstract level for which the appellant contends.

    Conclusions and orders

  63. The Court of Appeal reached the correct result on the matters of which the appellant complains in this Court. Further, the Court should not so modify the common law as to produce a different result.

  64. The appeal should be dismissed with costs.

    Hayne J & Bell J

  65. We agree with Kiefel J that, for the reasons her Honour gives, the appellant did not prove that the respondent's negligence was a cause of damage. We add only the following.

  66. For the purposes of the law of negligence, "damage" refers to some difference to the plaintiff. The difference must be detrimental. What must be demonstrated (in the sense that the tribunal of fact must be persuaded that it is more probable than not) is that a difference has been brought about and that the defendant's negligence was a cause of that difference. The comparison invoked by reference to "difference" is between the relevant state of affairs as they existed after the negligent act or omission, and the state of affairs that would have existed had the negligent act or omission not occurred: Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 181-182 [9].

  67. In this case, saying that a chance of a better medical outcome was lost presupposes that it was not demonstrated that the respondent's negligence had caused any difference in the appellant's state of health. That is, it was not demonstrated that the respondent's negligence was probably a cause of any part of the appellant's brain damage.

  68. As Gummow ACJ explains, to accept that the appellant's loss of a chance of a better medical outcome was a form of actionable damage would shift the balance hitherto struck in the law of negligence between the competing interests of claimants and defendants. That step should not be taken. The respondent should not be held liable where what is said to have been lost was the possibility (as distinct from probability) that the brain damage suffered by the appellant would have been less severe than it was.

  69. It may be that other cases in which it might be said that, as a result of medical negligence, a patient has lost "the chance of a better medical outcome" (for example, a diminution in life expectancy) differ from the present case in significant respects. These are not matters that need be further examined in this case. It need only be observed that the language of loss of chance should not be permitted to obscure the need to identify whether a plaintiff has proved that the defendant's negligence was more probably than not a cause of damage (in the sense of detrimental difference). The language of possibilities (language that underlies the notion of loss of chance) should not be permitted to obscure the need to consider whether the possible adverse outcome has in fact come home, or will more probably than not do so.

    Heydon J

  70. The trial judge, who conducted a lengthy trial with his customary balance and skill, correctly said that the case presented very complex issues. At the trial, one element of the plaintiff's case against the second defendant, who is the respondent in these proceedings (" the defendant" ), was that he had negligently ordered a lumbar puncture on 13 January 1991 which had caused brain damage. The trial judge rejected that case and it is not now pressed. Another element of the plaintiff's case was that the negligent failure of the defendant to ensure that a CT scan was carried out on 11 or 13 January more probably than not caused her brain damage. That case too was rejected by the trial judge and is not now pressed. That left the plaintiff with a contention that the failure to ensure a CT scan on 13 January deprived her of a chance, albeit a less than even chance, that the damage might have been avoided. There were other difficult issues facing the trial judge, with which this appeal is not concerned, about a failed case against another defendant, about whether the defendant was negligent at all, about how much of the plaintiff's damage was attributable to the defendant's negligence, and about the quantum of damages.

  71. The trial judge and the Court of Appeal accepted the factual aspect of the contention that the plaintiff had been deprived of a chance that the damage might have been avoided. But they were divided on a question of law, which the plaintiff put in this Court in the following way:

    Does the common law of negligence in Australia recognise a less than even chance of avoiding an adverse health outcome as an interest of value to a patient, the loss of which by reason of a doctor's negligence, can be compensated as damage suffered by that patient?

    The trial judge felt bound by an assumption as to the law in earlier authority to answer the question in the affirmative. The Court of Appeal upheld a challenge to that assumption and termed it "plainly wrong." It answered the question in the negative.

    A preliminary question

  72. Before that question of law can be answered in this appeal in relation to the plaintiff, it is necessary to conclude that the plaintiff did lose a less than even chance of avoiding an adverse health outcome as a result of the defendant's negligence. If she did not, the question does not arise.

    What was the defendant's negligence?

  73. The trial judge found that the defendant's negligence lay in his failure to order a CT scan shortly after the plaintiff's father drew a nurse's attention to the fact that the plaintiff, who was in hospital under the defendant's care, was staring and unresponsive at 11am on Sunday 13 January. Following observations by the nurse and a registrar, the defendant was summoned. He advised that a lumbar puncture be performed, as it was, but not a CT scan.

    What damage did the defendant's negligence cause?

  74. The damage which the plaintiff suffered was brain damage. Some of that damage was caused by increased intracranial pressure arising from a tumour which had been growing for over two years and from a build-up of cerebral spinal fluid ("hydrocephalus"). The occurrence of that damage manifested itself in symptoms – staring into space and a "seizure" – which were observed in the plaintiff from 11.45am on Monday 14 January. The trial judge put the same point from another angle in finding that "the failure to relieve the plaintiff's intracranial pressure during the twenty-four hour period prior to the plaintiff's decline on 14 January 1991 was causative of brain damage occurring at and following the time of the plaintiff's observed decline on 14 January." The trial judge found that that failure to relieve pressure had been caused by the defendant's failure to order a CT scan. The trial judge found that this brain damage was no greater than 25 percent of the plaintiff's total brain damage. The balance was caused by the tumour, the hydrocephalus, surgery on 16 January to remove the tumour, and subsequent medical treatment.

    The reasoning of the courts below

  75. Before 11am on 13 January the plaintiff was not thought to be suffering from a tumour. It is not disputed that if the CT scan had been called for soon after 11am on 13 January, the tumour and the hydrocephalus would have been discovered, an opportunity for treatment would have arisen, and that opportunity would have been taken pending surgery three days later. The crucial question is whether the chance of an occurrence of brain damage at 11.45am on 14 January could have been reduced if the defendant had arranged for a CT scan on 13 January. Despite the exceptional skill with which counsel for the plaintiff assembled and presented the arguments for an affirmative answer, the answer is in the negative.

  76. The reasoning of the courts below was as follows. If a decision to perform a CT scan had been made after 11am on 13 January, it would have required the summoning, from outside the hospital or from within the hospital after other medical activity had finished, of an anaesthetist, a radiologist, and a radiographer. It could have taken five to six hours to arrange and carry out the CT scan. But the CT scan would have detected the tumour and the hydrocephalus. The plaintiff would then have been treated with steroids rather than by the insertion of an intracranial drain to remove cerebral spinal fluid, because although Dr Maixner, a neurosurgery registrar on duty on 13 January (who later assisted Mr Johnston to remove the tumour), would have preferred the latter course, her superior, Mr Johnston, a neurosurgeon, would have adopted the former. That treatment would probably not have avoided the 25 percent share of the brain damage attributed to the negligence. But it would have created a less than even chance of avoiding that 25 percent share of the brain damage.

    The courts below diverge

  77. At this point the reasoning of the courts below diverged. The trial judge thought that the relevant chance was 40 percent. The Court of Appeal considered that the trial judge's figure was too high, and that the correct figure was not more than 15 percent. The Court of Appeal said:

    It is clear from his Honour's analysis that in reaching 40% he weighed in the scales the likely efficacy of a drain in reducing pressure. To do so, we think failed to give weight to the finding that is implicit in his reasons that on the balance of probabilities steroids would have been administered. If a loss of a chance is compensable, here, one cannot ignore that on the findings it is loss of a chance of a better outcome in circumstances where the administration of steroids was the proven likely treatment.

    (Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 555 [245]

  78. It is convenient now to proceed by inquiring whether the trial judge erred in relation to the insertion of a drain before 11.45am on 14 January, whether in any event the possibility that a drain could be inserted before 11.45am was legitimately open for consideration, and whether the insertion of a drain after 11.45am but two hours faster than it actually was would have mattered.

    Did the trial judge consider the possible insertion of a drain before 11.45am?

  79. Contrary to what the Court of Appeal said and the defendant submitted, on their true construction the reasons for judgment of the trial judge do not reveal the supposed error of finding that, while the likely treatment would have been steroids, there was a possibility of placing a drain before 11.45am, and that this possibility increased the chance of a favourable outcome. The trial judge did say that inserting an intracranial drain would have "given the plaintiff a chance of a better outcome than the prescription of steroids", but he put that consideration aside in view of the fact that the decision about what would have happened rested with Mr Johnston, and he preferred steroid treatment.

    Should the possible insertion of a drain before 11.45am have been considered?

  80. Although the trial judge did not commit the supposed error of taking into account the possibility that a drain might have been employed before 11.45am on 14 January, counsel for the plaintiff contended that it would not have been an error. He submitted that since the trial judge was dealing with hypothetical events, the trial judge was entitled to include the possibility of using a drain as one factor in his assessment. That submission must fail.

  81. The trial judge found that the treatment which would have been administered on 13 January to reduce pressure would have been "either .... drainage or prescription of steroids." These treatments are expressed as being alternative, not concurrent. The same is true of the following finding:

    I consider the use of a drain would have given the plaintiff a chance of a better outcome than the prescription of steroids, but in assessing the value of the lost chance I must heed the superior role Mr Johnston would have had in the decision-making process if he became the treating surgeon. Whichever of the two forms of treatment would have been undertaken, the plaintiff was deprived of the chance of having it by the ... defendant's breach of duty and was consequently deprived of the chance of a better outcome.

    Once the trial judge decided that Mr Johnston's preferred method of steroid treatment would have been the one embarked on, the trial judge saw the drain technique as an option only if the steroid treatment failed. In that respect the trial judge was accepting Mr Johnston's evidence. In particular, Mr Johnston said that the steroid treatment by itself would have continued until it was clear that "there was not improvement" or there was some deterioration in the plaintiff's condition. If either of those events happened, a drainage procedure would be employed – but only then. Mr Klug, a neurosurgeon who gave evidence for the plaintiff, considered that the steroid treatment would work within 24 hours. It would seem that if a CT scan had been performed within a five or six hour period after the incident at 11am on 13 January, and the decision to prescribe steroids had been taken some time after that, steroid treatment could not have commenced earlier than 6pm on 13 January, which would have given the steroids only 17 or 18 hours to take effect before 11.45am on 14 January. But Mr Klug did not say that steroid treatment could work within 17 or 18 hours. Hence a decision that it had not brought about an improvement could not have been made until after the episode on 14 January. And there was no deterioration until that episode. Accordingly there was no possibility, if Mr Johnston's approach was being followed, that a drainage procedure would have been employed and that this would have reduced the chance of the 14 January episode taking place. Once the trial judge decided that Mr Johnston's non-negligent decision would have been to treat the steroid path as an alternative to the drainage path and to pursue the former in the first instance, both he and the Court of Appeal were right to exclude the possibility that the drainage path had a chance of bringing about a better outcome – whether or not, as the trial judge thought, that chance was greater than the chance associated with steroid treatment.

  82. The plaintiff submitted that both steroid treatment and a drain insertion could have been used, and it was not an either/or choice. It is true that the trial judge stated that it might be necessary to insert the drain on 14 January, even if steroids had been prescribed the day before (above at [81]). But on the non-negligent approach which would have been followed, the necessity to insert the drain would only have arisen if there was some reason for thinking that the steroid treatment had failed or that the plaintiff's condition had deteriorated. There was no evidence that it would have been possible to say that the steroid treatment had failed until more than 24 hours had passed, and there was no deterioration in the plaintiff's condition until the episode at 11.45am on 14 January. However long the period over which the damage was suffered, the trial judge did not fix it as starting before 11.45am.

    Would the insertion of a drain after 11.45am have helped?

  83. The trial judge pointed out that if a CT scan had been carried out on 13 January, it would not have been necessary to perform the CT scan and the EEG on 14 January. Hence a drain could have been inserted after 11.45am on 14 January, approximately two hours earlier than it was. He thought that this would have increased the chance of a better outcome[9]. This reasoning rests on the idea that if the drain were inserted after the plaintiff's decline at 11.45am but before 3.10pm there was an increased chance of avoiding brain damage. This aspect of the trial judge's reasoning was supported by counsel for the plaintiff in this Court. He relied on Mr Johnston's evidence that if the plaintiff were on steroid treatment, she would have been closely monitored by neurological staff, that they could have been available for speedy intervention, and that once the episode of 14 January commenced there could have been speedy intervention by inserting a drain. This evidence was qualified to some degree by other evidence from Mr Johnston: the neurological staff were available anyway, and intervention would have been delayed by the need to arrange a theatre. But Mr Johnston accepted that time would have been saved if the CT scan had been carried out the day before.

  84. It is desirable to analyse the argument of counsel for the plaintiff by reference to alternative factual bases. One factual possibility is that the relevant brain damage suffered by the plaintiff took place in a relatively short period of time around 11.45am. The other factual possibility is that the brain damage took place over a continuous period or series of periods for some time after 11.45am. The trial judge said the brain damage occurred "at and following the time of the plaintiff's observed decline on 14 January." He thus fixed a time when the damage started. But he did not explicitly make a finding choosing between the two possibilities as to when it ended.

  85. If the matter is approached on the first possible factual basis, namely that the plaintiff's brain damage took place in a relatively short period of time around 11.45am, then inserting the drain two hours earlier than it actually was on 14 January would not have affected the occurrence of the damage suffered at 11.45am. It would not have increased the chance of that deterioration in the plaintiff's condition being avoided or reduced, and that is the vital inquiry. To that inquiry nothing that could only have happened after 11.45am on 14 January matters. Once the trial judge decided that it was Mr Johnston's course of treatment which would have been followed – steroid treatment but no drain insertion until either the steroid treatment was seen to be failing or the plaintiff's condition deteriorated – there was no occasion before the actual deterioration in the plaintiff's condition at 11.45am on which to decide to employ a drain. The trial judge spoke of the elimination of the delay for the CT scan and the EEG as permitting the earlier insertion of the drain and as therefore increasing the chance of a better outcome "[i]f, notwithstanding the prescription of steroids on 13 January, the decline had still occurred on 14 January." But there was no evidence or finding that any decline on 14 January took place until 11.45am, nor any evidence or finding that an occasion arose before 11.45am on 14 January on which the effectiveness of the steroid treatment might be reviewed and a decision made to insert a drain. Hence any saving in time in employing a drain by reason of what ought to have happened the day before was immaterial.

  86. If the matter is approached in the light of the second factual possibility – that the relevant brain damage occurred over a continuous period or series of periods beginning at 11.45am – the events in the afternoon, and the saving of time had the CT scan taken place on 13 January, could have significance. The problem is that there was no evidence, and the trial judge was therefore not able to make any finding, about whether the occurrence of brain damage was a continuing process or about whether an insertion of the drain two hours earlier than it was inserted would or could have avoided or reduced the continuing process of brain damage.

  87. The trial judge accepted, and the plaintiff relied on, some general medical evidence that the longer the delay between deterioration and intervention by neurosurgeons the greater the likely damage. But this evidence does not overcome the difficulty that there was no occasion of deterioration to suggest the need for intervention by a neurosurgeon in the form of adopting the drain technique before the damage began to be suffered at 11.45am. Nor does it overcome the difficulty that it has not been shown that in the specific case of this plaintiff the two hours would or could have mattered thereafter.

    Did the failure to adopt steroid treatment destroy any chance of avoiding the brain damage that happened on 14 January?

  88. The trial judge found that whichever of the two forms of treatment would have been undertaken if a CT scan had revealed the tumour and the hydrocephalus – prescription of steroids or drainage treatment – the plaintiff was deprived of the chance of having it and was consequently deprived of the chance of a better outcome. The chance of a better outcome arising from drainage treatment has been discussed in the preceding paragraphs, and excluded. What of the chance of a better outcome arising from the preferred form of treatment, steroid treatment?

  89. Apart from Mr Johnston's last two answers in re-examination (see Gummow ACJ at [43]), Mr Johnston gave other relevant evidence. In answer to the question whether steroid treatment "would have reduced the risk of, if not avoided, the incident that occurred on the 14th", Mr Johnston said: "I'm not sure." After a debate about an objection, he continued his answer as follows:

    [S]teroids are not particularly effective in, say, situations where the pressure is due to hydrocephalus. They are more effective where there is actual brain swelling, brain oedema. So I don't know – I can't say that steroids would have had a very significant effect, and I certainly can't say – let us assume that she had been on steroids from Sunday at 11am – that that would have stopped the episode on Monday. I think that's not a reasonable supposition.

    In a later answer he said that after initiating the steroid treatment:

    it is problematic what would then have happened. I mean, it is entirely possible that actually the same course of events would have happened, that we would have been closely monitoring her and then on the Monday she would have had the deterioration.

    In re-examination he was asked how steroid treatment would have assisted the plaintiff on 13 January. He answered that he thought that steroids were not particularly effective in relation to hydrocephalus, but were more effective where there was brain swelling or tumour-related swelling, and could have reduced that swelling. Although it follows from the trial judge's findings that steroid treatment would have commenced no earlier than 6pm on 13 January, the trial judge did not make findings, because the evidence did not permit him to make findings, about the time by which steroid treatment might have become effective or the time at which or the circumstances in which the effectiveness of steroid treatment might have been reviewed and steps taken to insert a drain.

  90. The plaintiff's argument in relation to the effect of the steroid treatment by itself came to this. The plaintiff accepted that Mr Johnston testified that the proposition that the steroid treatment would have prevented the 14 January episode was "not a reasonable supposition", "would be absolutely a guess", and was "entirely speculative." But the plaintiff submitted that this evidence was directed to whether it was more probable than not that the steroid treatment would have prevented the episode. That was not, for present purposes, the relevant question. The relevant question was whether there was a chance that they may have done so. The plaintiff pointed to evidence from Mr Johnston that steroids could have beneficial effects even if they would not have prevented the 14 January episode. The problem is that Mr Johnston's evidence must be taken as a whole. His last answer in re-examination (see Gummow ACJ at [43]) was that while "you could certainly make an argument" that the use of steroids "may have prevented the episode" [emphasis added], it was not an argument one could have confidence in: "nobody could answer that with any certainty." He thought that it "would be absolutely a guess", "entirely speculative" and "very speculative" whether steroid treatment would have created a chance of avoiding the 14 January incident. That is, it was not possible to say that there was even a chance that steroid treatment may have prevented the episode. If Mr Johnston's evidence had stood alone, it would not have been right to conclude that the lost chance of a better outcome was quantifiable. The Court of Appeal correctly said of his evidence:

    If that was the only evidence, the [plaintiff] would not have established that she was entitled to an award of damages for the loss of a chance of a better outcome as the evidence would not have [risen] above there being a speculative chance.

    Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 554 [241]

  91. The Court of Appeal, however, then referred to certain evidence of Mr Klug. It involves three propositions. First, Mr Klug said that if a CT scan had been performed on either 11 January or 13 January it "would have indicated the presence of hydrocephalus" in addition to the existence of the tumour. Secondly, he said that if the plaintiff was not in extremis (which the plaintiff was not on 13 January) he:

    would have used high dose corticosteroids in the first instance which is very effective in improving the situation of children with this disorder. I would have then planned when to carry out a definitive procedure to remove the tumour which would have involved, as part of the removal of the tumour, also the insertion of an intraventricular drain. If the child is in good condition I would see no reason to, as a preliminary, insert a ventricular drain.

    He said that the administration of high dose corticosteroids "invariably, within 24 hours, would lead to substantial improvement in the condition and enable one to more fully assess the patient and plan on a semi-elective basis to undertake the operation." Thirdly, he said that on 13 January he would have had:

    to very carefully assess the condition of the child at that time and it could well be, if I thought the hydrocephalus was extreme and causing or having the risk of potential serious deterioration, I may have used a combination of treatment at that time, namely, a ventricular drain and corticosteroids. I think it's a delicate decision here, one would have to very carefully analyse the situation of the child from a neurologic point of view.

  92. The Court of Appeal said of the third proposition in this evidence, ibid at 554 [242]:

    The reasonable inference from this evidence is that even if there were hydrocephalus, steroids would still have some effect, though the evidence did not permit any conclusion as to what that effect would have been.

    Nor did the evidence permit any conclusion that the effect would have been to reduce the chance of the occurrence of the brain damage which took place at or after 11.45am on 14 January. In these circumstances the Court of Appeal's conclusion that the plaintiff "lost some chance of a better outcome which ranged between speculative and some effect" does not follow, ibid at 554-555 [243].

  93. The damage on 14 January was caused by the intracranial pressure generated by the tumour and the hydrocephalus in combination. The steroid treatment which Mr Johnston would have adopted would not have been "particularly effective" where the pressure was due to the hydrocephalus, but was more likely to be effective where there was actual brain swelling whether caused by the tumour or otherwise. There is no evidence as to what role the tumour actually played in causing brain compression or a vascular compromise, independently of the proposition that it operated in combination with the hydrocephalus. As explained above (at [81]), the period from when the steroid treatment would have commenced, ie 6pm on 13 January, until the episode at 11.45am on 14 January was less than 18 hours. While Mr Klug's experience indicated that the steroid treatment would "invariably" bring about improvement within 24 hours, he did not give evidence about whether it could be expected within 17 or 18 hours. Nor did he give evidence about whether his conclusions drawn from general experience necessarily applied in the circumstances of the plaintiff: the trial judge found that her tumour had existed for 700 to 945 days, was 3.5cm in diameter, was "extensive" and was "very advanced."

  94. All these factors support Mr Johnston's view that it "would be absolutely a guess", "entirely speculative" and "very speculative" whether steroid treatment would have created a chance of avoiding the 14 January incident. It follows that the conclusions in the courts below that there was a chance of avoiding the brain damage on 14 January had the defendant arranged a CT scan soon after 11am on 13 January cannot be sustained.

    The Court of Appeal's lack of confidence

  95. The Court of Appeal's selection of a 15 percent figure for the lost chance was very tentative. Their Honours said that if they "were forced" to place a percentage figure on the lost chance they would be "loathe to assess it as greater than 15%", and said that that figure was "at most" that which reflected the lost chance: Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 555 [245].

  96. Further, the difference between 40 percent as found by the trial judge and 15 percent is a large one. One number is nearly three times the size of the other, and the financial consequences are significant. The difference is actually larger because of the Court of Appeal's doubts about whether even 15 percent was correct. But the very fact of this large difference between conclusions reached after such careful consideration in each of the courts below of a very difficult case supports the conclusion that it was in truth not possible to arrive at any conclusion on the question of whether the negligence caused the plaintiff to lose a chance of avoiding or reducing the damage. The difference suggests that the condition of the evidence left them no alternative but to grope towards speculative outcomes which it was impossible for them firmly to grasp.

    Conclusion

  97. The question of law which the plaintiff wishes to agitate was argued by the parties fully and forcefully because on one factual basis it would have been a live and decisive issue. But now the question has ceased to be live and decisive. The question has become hypothetical in the sense that the assumption it rests on has turned out to be incorrect. The question is controversial among lawyers and in other cases, but as between the plaintiff and the defendant in this case the controversy has turned out to lack concreteness. For them it has become moot. There is no answer to the question posed which will produce any consequences for the parties. The question has become purely abstract and academic. The only significance of an answer would lie in what future courts would make of it. They are likely to treat it not as a decision, but only as a dictum; not as the resolution of a controversy, but only as advice; not as an event, but only as a piece of news.

  98. The consciousness of parties and their legal representatives that the outcome of a debate about the correctness of contested propositions of law is decisively important to the interests of those parties often greatly assists the sharpness and quality of that debate. Doubtless it did so here. But the efficacy of a debate does not depend only on whether the participants in the debate have that consciousness. The efficacy of its resolution depends on the court sharing that consciousness and being assisted by that consciousness. Here a stage has been reached in a journey along the path to decision which has caused that consciousness to cease to exist because an issue has ceased to be decisively important. No assistance can be gained from a consciousness that has ceased to exist. In this field, for me, at least, to embark on difficult and doubtful inquiries in an attempt to answer the question without the assistance to be gained from that consciousness is a potentially very dangerous course. This is a case in which, since it is not necessary to do so, it is desirable not to.

  99. Accordingly, the appeal must be dismissed.

    Crennan J

  100. This appeal arises out of an action in medical negligence, and turns on the application of fundamental principles of causation in accordance with which the appellant failed to prove that the respondent's negligence caused or contributed to cause damage. The facts are set out in the reasons of Gummow ACJ and Kiefel J. The appellant raises the question of whether Australian law does or should permit recovery of damages where the breach of a duty of care results in the loss of a chance of a better medical outcome, where the chance of avoiding certain damage which occurred was assessed by the trial judge at 40 per cent: Tabet v Mansour [2007] NSWSC 36 at [378], [434]. That question should be answered in the negative for the reasons given by Kiefel J, with which I agree.

  101. As recognised by her Honour and also by Gummow ACJ and Hayne and Bell JJ, the adoption in personal injury cases of "loss of a chance" as a basis for liability would represent a major development in the common law. If the appellant's arguments were accepted, the respondent would be held liable where what had been lost was the possibility (but not the probability) that the brain damage suffered by the appellant would have been less catastrophic than it was.

  102. The present requirement of proof of causation in personal injury cases results in boundaries being drawn which differ from those which are relevant to liability for pure economic loss. Policy considerations which tell against altering the present requirement of proof of causation in cases of medical negligence include the prospect of thereby encouraging defensive medicine, the impact of that on the Medicare system and private medical insurance schemes and the impact of any change to the basis of liability on professional liability insurance of medical practitioners. From the present vantage point, the alteration to the common law urged by the appellant is radical, and not incremental, and is therefore the kind of change to the common law which is, generally speaking, the business of Parliament.

  103. I agree with the orders proposed by Gummow ACJ and Kiefel J that the appeal should be dismissed with costs.

    Kiefel J

  104. Reema Tabet (" the appellant" ) was six years old when she was readmitted to hospital on 11 January 1991 with symptoms of vomiting and headaches. She had recently suffered from varicella (chickenpox). A CT scan taken on 14 January revealed that she had a large brain tumour. Differing opinions were given by expert witnesses as to whether a scan should have been ordered by the respondent, a specialist paediatrician, at an earlier point in time, given the symptoms exhibited by the appellant. The trial judge, Studdert J of the Supreme Court of New South Wales, was persuaded that one was necessitated immediately after the appellant had been observed to be staring and unresponsive on the morning of 13 January, ibid at [193]. The finding of a negligent omission, on the part of the respondent, is not in issue on this appeal.

  105. The appellant suffered brain damage as a result of a neurological event which occurred on 14 January and which led to the CT scan being performed. Studdert J found that the damage was associated with intracranial pressure, produced by the pressure of the tumour and an excess of spinal fluid in the cranial cavity (hydrocephalus). That damage contributed to the severe, irreversible brain damage and consequent disability which the appellant now suffers. The other contributors were the tumour itself, the operation undertaken in an attempt to remove it and the treatment which followed, ibid at [382]. Studdert J attributed 25 per cent of the appellant's overall disability to that neurological event.

  106. Studdert J was not persuaded, on the balance of probabilities, that if the respondent had ordered a CT scan on 13 January and the appellant was treated upon the discovery of the tumour, such brain damage as occurred on 14 January would have been avoided. Her claim that such damage was caused by the respondent therefore failed. However, his Honour considered that she had been deprived of the chance of a better outcome by reason of the delay in the treatment she could have received and was entitled to be compensated for that loss. Earlier detection of the tumour would have enabled treatment, most probably by corticosteroids, in an attempt to reduce the intracranial pressure. This would have had some beneficial effect, his Honour held. His Honour assessed "the chance of a better outcome, and of avoiding the brain damage that occurred on 14 January 1991" at 40 per cent, ibid at [434]. His Honour applied that percentage to the figure representing the contribution of the event of 14 January to the appellant's overall disability in arriving at an award of $610,000.

  107. The Court of Appeal of the Supreme Court of New South Wales considered that, were damages to be assessed for the loss of the chance of a better outcome, they should be reduced to 15 per cent: Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 per Allsop P, Beazley and Basten JJA. The trial judge's assessment of the chance of a better outcome at 40 per cent took into account that an intraventricular drain would also have been inserted, thereby increasing that chance. The Court of Appeal held that this was contrary to the finding that treatment by corticosteroids rather than the placement of the drain was most likely to have been pursued. ibid at 555 [245]. However, the Court allowed the respondent's appeal and dismissed the claim. In its opinion to permit recovery for the claim for the loss of the chance involves a proposition which would revolutionise proof of causation of injury, ibid at 586 [381]. A decision of that Court which had adopted a loss of chance analysis (Rufo v Hosking [2004] NSWCA 391; (2004) 61 NSWLR 678 (and also Gavalas v Singh [2001] VSCA 23; (2001) 3 VR 404)) was considered by the Court of Appeal to have departed from conventional principles and it declined to follow it: Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 587 [389].

    Damage and causation in an action for medical negligence

  108. The three elements of a cause of action in medical negligence, necessary to be established in order to recover compensation, are a duty owed by the medical practitioner to the plaintiff to avoid harm which is reasonably foreseeable, a breach of that duty and damage which results from that breach. It is the third element which is the focus of this appeal. It incorporates both the fact of loss or damage having been suffered and the cause of that damage being the medical practitioner's negligent act or omission. Those facts are ordinarily required to be proved to the general standard, on the balance of probabilities: Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 at 355.

  109. Damage is an essential ingredient in an action for negligence; it is the gist of the action: Williams v Milotin (1957) 97 CLR 465 at 474. The action developed largely from the old form of action on the case, in which it was the rule that proof of damage was essential to a plaintiff's case[10]. In Brunsden v Humphrey (1884) 14 QBD 141 at 150 Bowen LJ pointed out that in certain classes of case the mere violation of a legal right imports damage, but that principle was "not as a rule applicable to actions for negligence: which are not brought to establish a bare right, but to recover compensation for substantial injury." Generally speaking "there must be a temporal loss or damage accruing from the wrongful act of another, in order to entitle a party to maintain an action on the case." [11] Negligence in the abstract will not suffice: Hay or Bourhill v Young [1943] AC 92 (at 116 per Lord Porter; Haynes v Harwood [1935] 1 KB 146 at 152).

  110. An action in negligence, said Bowen LJ, "is based upon the union of the negligence and the injuries caused thereby, which in such an instance will as a rule involve and have been accompanied by specific damage": Brunsden v Humphrey (1884) 14 QBD 141 at 150. Nevertheless the action on the case has itself been described as sufficiently flexible to enable judges to extend it to cover situations where damage was suffered in circumstances which called for a remedy: Sidaway v Governors of Bethlem Royal Hospital [1985] AC 871 at 883 per Lord Scarman. The Court of Appeal in this case observed that the common law has adapted to recognise different kinds of harm. But nowhere is it suggested that the requirement for damage itself can be dispensed with. Liability based upon breach of duty of care without proven loss or harm will not suffice.

  111. The common law requires proof, by the person seeking compensation, that the negligent act or omission caused the loss or injury constituting the damage. All that is necessary is that, according to the course of common experience, the more probable inference appearing from the evidence is that a defendant's negligence caused the injury or harm. "More probable" means no more than that, upon a balance of probabilities, such an inference might reasonably be considered to have some greater degree of likelihood; it does not require certainty: Bradshaw v McEwans Pty Ltd (1951) 217 ALR 1 at 6.

  112. The "but for" test is regarded as having an important role in the resolution of the issue of causation, although more as a negative criterion than as a comprehensive test: March v Stramare (E & M H) Pty Ltd (1991) 171 CLR 506 (at 515-516 per Mason CJ, 522 per Deane J). The resolution of the question of causation has been said[12] to involve the common sense idea of one matter being the cause of another. But it is also necessary to understand the purpose for making an inquiry about causation[13] and that may require value judgments and policy choices[14].

  113. Once causation is proved to the general standard, the common law treats what is shown to have occurred as certain[15]. The purpose of proof at law, unlike science or philosophy, is to apportion legal responsibility: March v Stramare (E & M H) Pty Ltd (1991) 171 CLR 506 at 509 per Mason CJ. That requires the courts, by a judgment, to "reduce to legal certainty questions to which no other conclusive answer can be given." [16] The result of this approach is that when loss or damage is proved to have been caused by a defendant's act or omission, a plaintiff recovers the entire loss (the "all or nothing" rule).

    The appellant's problem in proof of causation of physical damage

  114. In actions involving medical negligence the loss or damage claimed to have been suffered is ordinarily physical or mental injury or harm. When such injury or harm is proved the question then is whether it was caused by the negligent act or omission, such as a failure to diagnose or treat the disease or other condition from which the plaintiff then suffered. The difficulty which the appellant faced in this case was that the expert medical evidence did not establish the link between the omission of the respondent, with the consequent delay in treatment, and the brain damage which occurred on 14 January, necessary for a finding of causation. There was no evidence as to what harm might have been caused by the delay[17]. It could not be said that "but for" the delay the appellant would not have suffered brain damage. It follows from Studdert J's findings that the probability was that the tumour would have caused it in any event.

  115. Studdert J found that the best outcome for the appellant required that the intracranial pressure on the brain be relieved as soon as possible. There were two possible treatments available: the administration of corticosteroids or the insertion of an intraventricular drain, but his Honour made a finding that it was "more likely that steroids would have been prescribed rather than the placement of a drain": Tabet v Mansour [2007] NSWSC 36 at [378]. His Honour made no finding as to what the outcome of such treatment in the period of delay would have been. His Honour considered that corticosteroids would have had "some beneficial effect." The evidence did not permit a more specific or certain finding.

  116. The evidence as to the effectiveness of corticosteroids, to prevent the brain damage from occurring the next day, was limited. Of the many medical witnesses who gave evidence, only the evidence of two neurosurgeons touched upon the question. Mr Johnston agreed to the general propositions that the earlier the intervention the better the likely result and that corticosteroids would have improved the appellant's neurological condition. But more specifically, and in relation to the appellant's condition, he said that corticosteroids are not particularly effective in situations where the pressure is due to hydrocephalus and he could not say that if she had been given corticosteroids when her condition was first noted as deteriorating on 13 January, they would have stopped the neurological event occurring the next day. He said that it was "problematic" as to what would have happened if corticosteroids had been initiated and that "it is entirely possible that actually the same course of events would have happened." The use of steroids in avoiding the incident of 14 January he considered to be "entirely speculative." He did not think they would have been sufficiently effective, but he did not know.

  117. Mr Klug said that in non-acute situations high dose corticosteroids were "very effective in improving the situation of children with this disorder", but did not elaborate further.

  118. This evidence does not support a finding that any chance of a better outcome was as high as 40 per cent. The Court of Appeal observed that, whilst it might be inferred from this evidence that corticosteroids might have some effect, it "did not permit any conclusion as to what that effect would have been.": Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 554 [242]. It considered that, at the most, it could be said that the appellant "lost some chance of a better outcome which ranged between speculative and some effect" (ibid at 554-555 [243]), but went on to hold that even so, to permit recovery for the deprivation of the possibility, but not the probability, of a better outcome would be to significantly alter the existing law as to proof of causation of injury, in particular by redefining what is "harm", ibid at 586 [381].

    The appellant's arguments

    Redefining damage?

  119. The question raised by this appeal is whether the common law of Australia should recognise the loss of a chance of a better outcome, in cases where medical negligence has been found, as actionable damage. In Gregg v Scott Lord Nicholls of Birkenhead observed that it is "a question which has divided courts and commentators throughout the common law world" : [2005] UKHL 2; [2005] 2 AC 176 at 180 [1]. The same observation may be made with respect to civil law systems. In Australia the question has been considered in decisions of intermediate courts of appeal in addition to that the subject of this appeal[18]. It has only been touched upon in decisions of this Court: Chappel v Hart (1998) 195 CLR 232; Naxakis v Western General Hospital (1999) 197 CLR 269.

  120. The argument for the appellant, for the acceptance by this Court of the loss of a chance of a better outcome as damage, seeks to draw support from the approach taken by courts of some common law countries, notably the United States of America, and some civil law countries, in particular France, as relevant to what is submitted to be the choice now presented. It is not suggested that a review of other legal systems reveals that there is a correct solution. So much may be accepted. Decisions by courts of other countries, including common law countries, concerning cases of this kind are made in the framework of their substantive law, the principles and policies which inhere in it and the requirements for proof of causation and damage which may or may not be adaptable to accommodate such a claim.

  121. In argument for the appellant it was stressed that the only change necessary to accommodate a loss of chance claim is to the type of harm or damage which may result in medical negligence cases. The shift from physical harm to the chance of a better outcome as representing loss is said not to alter or contradict the requirement of proof of loss or damage on the balance of probabilities. But it is accepted by the appellant that the reformulation of the damage may affect the causation question by "shaping" it[19].

    Analogy with loss of commercial opportunity cases

  122. It was argued that the loss of an opportunity of a better outcome in a patient's illness or condition should not be seen as novel. The law in Australia already recognises the loss of a commercial opportunity as actionable damage. Accepting that there is a commercial interest in realising an opportunity, it was submitted for the appellant that a person likewise has an interest in their medical outcome.

  123. It was recognised in Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 that a loss of the opportunity to obtain a commercial advantage or benefit is loss or damage for the purposes of s 82(1) of the Trade Practices Act 1974 (Cth), where the cause of action arose under s 52(1) of that Act. Previous decisions allowing for recovery had been based in contract, where the breach of the promise to provide the chance itself gave rise to the loss of that chance: Chaplin v Hicks [1911] 2 KB 786; Fink v Fink (1946) 74 CLR 127. But as Brennan J said, in cases under s 82(1), "as in cases of tort where damage is the gist of the action, a lost opportunity may or may not constitute compensable loss or damage" and it must be proved in some other way: Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 at 359.

  124. What cases in contract, such as The Commonwealth v Amann Aviation Pty Ltd (1991) 174 CLR 64 and Sellars v Adelaide Petroleum NL, have in common is that the commercial interest lost may readily be seen to be of value itself. The same cannot be said of a chance of a better medical outcome or a person's interest in it. Lord Hoffmann observed in Gregg v Scott [2005] 2 AC 176 at 197 [83] that most cases where there has been recovery for loss of a chance have involved financial loss, where the chance itself can be regarded as an item of property. And in Sellars v Adelaide Petroleum NL Brennan J observed that, "[a]s a matter of common experience, opportunities to acquire commercial benefits are frequently valuable in themselves." So long as an opportunity provides a substantial and not merely a speculative prospect of acquiring a benefit, it can be regarded as of value and therefore loss or damage: Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 at 364. A loss of a chance of a better medical outcome cannot be regarded in this way. As the assessment of damages in this case shows, the only value given to it is derived from the final, physical, damage.

    Loss of chance as damage – civil law countries

  125. In submissions for the appellant support was sought from the recognition given by other countries, in particular France, to lost chances as a kind of harm[20]. But the general clauses of the Code Civil (arts 1382 and 1383) in no way explain what is meant by "dommage" [21] and the approach taken by French courts to the question of damage is different. Professor Markesinis suggests that the tendency towards generalisation and abstraction is most evident in the way French judges deal with the issue of damage. He says that they look for general elements such as certainty of damage and the legitimacy of the interest affected, rather than by finding the form taken by the harm in the case in question[22]. The courts typically regard the loss of a chance as certain if it is real and not just hypothetical[23].

  126. An approach closer to that taken in cases in Australia involving loss of commercial opportunity appears to be adopted in civil law countries which do not favour the loss of chance theory[24]. Some do not regard chances, even the chance to be cured, as chances which can be valued on their own[25]. In these countries the possibility of defining a loss of a chance as damage is given only if the chance has an economic value and is accepted as an interest the law will protect, for example a legally recognised opportunity to yield a profit[26]. If it has no economic value it cannot qualify as damage which can be compensated.

    Loss of chance as independent harm

  127. The approach of the French courts was also relied upon to support the submission that loss of a chance in medical negligence cases may be seen as a kind of harm independent of the physical harm occasioned. It would strengthen the argument for acceptance of such harm as damage if it were shown to have a separate, independent existence[27]. Professor Khoury says that some judgments of the French courts acknowledge the uncertainty inherent in the loss of chance in medical negligence cases but justify compensation on the basis that it is a loss independent of the final damage[28]. Despite some strong criticisms, the acceptance of loss of chance cases continues to be supported by the courts. Loss of chance is said to be the "preferred tool" for dealing with causal difficulties created by scientific uncertainty[29]. This suggests that a policy choice is involved.

  128. There is a real question in this case whether the loss of a chance of a better outcome could be said to be independent of the physical harm suffered by the appellant. Professor Khoury refers to criticism levelled by French commentators at loss of chance in cases of this kind as an alternative head of injury. They suggest that it is not truly distinct, for its calculation is always contingent upon applying the percentage of the lost chance to the quantum of damage relating to the final injury[30]. It will be recalled that that is the process which was undertaken by the trial judge in this case. The commentators argue that when the final injury occurs it "absorbs" the intermediate damage, which loss of chance represents, so that when damages for a lost chance are granted, they constitute, in effect, partial compensation for the actual injury[31]. This point was not lost on the Court of Appeal in this case: Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 584 [371] and 585 [375].

  129. In her article "Causation and Risk in the Highest Courts of Canada, England, and France" ((2008) 124 Law Quarterly Review 103) Professor Khoury discusses whether an increase in the risk of injury, there being some cases of lost chance which come within this description, might be considered an independent head of damage. It is not necessary to further consider this question on this appeal. Although the Court of Appeal expressed the view that the "so-called loss of an opportunity" was in reality a claim based upon an increased risk of harm (ibid at 585 [377]), and recovery therefore not permitted[32], it was not discussed in the present context.

  130. Professor Khoury suggests that if loss of chance were a truly independent type of injury, defendants would be forced to compensate the plaintiff even if the lost chance resulted in no actual injury[33]. The same point is made by some German commentators. The example given in the Digest of European Tort Law[34] is where there is a delay in treatment which is said to reduce the chances of being healed by 40 per cent, but by the time of trial the claimant has been healed. Because the loss of the chance is the relevant damage, in theory a claim may be made as soon as the chance is lost or reduced.

  131. Another aspect of the problem identified may be seen in Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176, where, however, the chance had not played out. The statistical model relied upon by the plaintiff gave his chances of survival for 10 years at the time he consulted with the defendant, who failed to treat his tumour, as 42 per cent. The plaintiff was still alive at trial, when his chances were then assessed at 25 per cent, and he was still alive when the appeal was heard. Considerable uncertainty attended the question as to what his chances were. As Lord Phillips of Worth Matravers MR observed, statistically his prospects of surviving had been improving up to trial and were increasing daily thereafter. The model was inadequate to provide a conclusion as to his chances: Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 216-217 [156]-[157]. By the time of the appeal it was not possible to reach the trial judge's conclusion. The likelihood that the delay in treatment had any effect diminished the longer the plaintiff survived, ibid at 220 [169].

    Matsuyama v Birnbaum 890 NE 2d 819 (Mass 2008)

  132. The statistical evidence in that case was regarded by Marshall CJ[35] as wholly reliable: Matsuyama v Birnbaum 890 NE 2d 819 at 833-834 (Mass 2008). Mr Matsuyama died after his gastric cancer metastasised to an inoperable state. The evidence of an expert gastroenterologist was that the cancer would have been diagnosed had there been appropriate testing when Mr Matsuyama consulted the defendant and he might have been capable of cure if treated then, ibid at 826 (Mass 2008). He gave evidence of the method employed by oncologists of classifying cancer of this type into stages, with each stage signalling a more advanced cancer and carrying a statistically diminished chance of survival, ibid at 826 (Mass 2008). The jury found that Mr Matsuyama had a 37.5 per cent chance of survival at the time he saw the defendant, ibid at 828 (Mass 2008). There was no evidence of this kind in the present case.

  133. The starting point for the method given by Marshall CJ, to be employed in such cases, is to calculate the total damages which might be allowed for the death or personal injury. The person's present chance, expressed as a percentage, of survival or cure, is deducted from the chance they had immediately prior to the act of negligence and then applied to the figure for total damages, ibid at 840 (Mass 2008).

  134. It was argued for the appellant that the award of "proportional damages", made in Matsuyama v Birnbaum, offers a workable solution to cases of this kind and is consistent with Australian authority as to the assessment of past hypothetical situations in the assessment of damages.

  135. It is important to bear in mind, in connection with this aspect of the appellant's argument, the distinction between the loss or damage necessary to found an action in negligence, which is the injury itself and its foreseeable consequences, and damages, which are awarded as compensation for each item or aspect of the injury: Mahony v J Kruschich (Demolitions) Pty Ltd (1985) 156 CLR 522 at 527.

  136. Different standards apply to proof of damage from those that are involved in the assessment of damages. Sellars v Adelaide Petroleum NL (1994) 179 CLR 332 confirms that the general standard of proof is to be maintained with respect to the issue of causation and whether the plaintiff has suffered loss or damage (at 355 per Mason CJ, Dawson, Toohey and Gaudron JJ and 367 per Brennan J). In relation to the assessment of damages, as was said in Malec v J C Hutton Pty Ltd (1990) 169 CLR 638 (at 643 per Deane, Gaudron and McHugh JJ), "the hypothetical may be conjectured." The court may adjust its award to reflect the degree of probability of a loss eventuating. This follows from the requirement that the courts must do the best they can in estimating damages; mere difficulty in that regard is not permitted to render an award uncertain or impossible: The Commonwealth v Amann Aviation Pty Ltd (1991) 174 CLR 64 at 83 per Mason CJ and Dawson J, citing Fink v Fink (1946) 74 CLR 127 at 143.

  137. Thus in the case of the loss of a commercial opportunity, the plaintiff must first establish the fact of the loss, for example by reference to the fact that it had a commercial interest of value which is no longer available to be pursued because of the defendant's negligence. The damages assessed of that loss, the estimation of its value, reflect the chance, often expressed in a percentage, that the opportunity would have been pursued to a successful outcome. The award is proportionate in that sense.

  138. The "proportional damages" awarded in Matsuyama v Birnbaum do not involve such an assessment. The damages are expressed as a proportion of the total damages which might have been awarded for Mr Matsuyama's wrongful death, but for which the defendant could not be held liable. They have the effect of providing for proportionate liability.

  139. The jury in Matsuyama v Birnbaum 890 NE 2d 819 (Mass 2008) had not identified the loss of the chance as damage. Their finding of causation was that the defendant's negligence was a "substantial contributing factor" to Mr Matsuyama's death and they awarded 37.5 per cent of damages for wrongful death: ibid at 828. That outcome was maintained on appeal, but the test of causation was corrected, the "but for" test being held to apply, ibid at 842. But in applying that test the Court appears to have focussed upon only Mr Matsuyama's former chance of survival as the fact relevant to the assessment of his position. It does not appear to have considered the fact that he had cancer, which gave the probability that he would not survive, as relevant. As Gonthier J said, giving the judgment for the majority of the Supreme Court of Canada (on appeal from the Court of Appeal for Quebec) in Laferrière v Lawson [1991] 1 SCR 541 at 591, consideration of the entire factual situation is necessary on the issue of causation.

    Causation in this case

  140. The issue whether damage has been caused by a negligent act invites a comparison between a plaintiff's present position and what would have been the position in the absence of the defendant's negligence: Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 182-182 [9] per Lord Nicholls of Birkenhead; Harriton v Stephens (2006) 226 CLR 52 at 104 [167] per Hayne J. Such an inquiry directs attention to all the circumstances pertaining to the plaintiff's condition at the time he or she sought the medical treatment which was not properly provided. The question of whether harm or damage has been suffered is bound up in the question of causation.

  141. In the present case the appellant suffers from severe brain damage, some of which occurred on 14 January 1991. It is that damage which is the focus of the inquiry about causation. At the time a CT scan should have been performed she had a large brain tumour which was causing intracranial pressure. Unrelieved it was almost certainly going to cause the brain damage which eventuated. A conclusion that earlier treatment would have altered that outcome is not possible. It could not therefore be demonstrated that the respondent's negligence was probably a cause of the appellant's brain damage.

  142. Expressing what is said to have been lost as the loss of a chance was said by Gonthier J in Laferrière v Lawson to divert attention from the proper connection between fault and damage. It is artificial and breaks the causal link: Laferrière v Lawson [1991] 1 SCR 541 at 591. I respectfully agree. One commentator to whom his Honour referred[36] suggests that in cases of the kind in question what is involved is in truth not a loss of a chance. The factors present in that chance have played themselves out when physical injury or death occurs. What is in issue is a past event. It was to this opinion that Gaudron J referred, with approval, in Naxakis v Western General Hospital (1999) 197 CLR 269 at 280 [36].

    The standard of proof

  143. Resort to the language of "chance" cannot displace the analysis necessary for the determination of the issue of causation of damage. Properly analysed, what is involved in the chance referred to in this case is the possibility, to put it at its highest, that no brain damage would occur or that it would not be so severe. They are the "better medical outcomes" involved in the chance. Expressing what is said to be the loss or damage as a "chance" of a better outcome recognises that what is involved are mere possibilities and that the general standard of proof cannot be met. Thus the appellant could only succeed if the standard of proof is lower than the law presently requires.

  144. Gregg v Scott confirmed for the United Kingdom that the general standard of proof should be maintained with respect to claims for damages for medical negligence. Lord Nicholls was the only member of the House of Lords to consider that the law should not require proof on the balance of probabilities and should recognise a person's prospects of recovery as real: Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 189 [42]. The Supreme Court of Canada in Laferrière v Lawson confirmed that if a case did not meet the test of causation applying the general standard of proof, then recovery should be denied: Laferrière v Lawson [1991] 1 SCR 541 at 608 per Gonthier J.

  145. The general standard of proof required by the common law and applied to causation is relatively low. It does not require certainty or precision. It requires that a judge be persuaded that something was probably a cause of the harm the plaintiff suffered. Historically the standard may have been chosen in order to minimise errors in civil jury trials[37], but it nevertheless serves also to accommodate a level of uncertainty in proof.

  146. In countries like France evidence must approach certainty in proof of causation[38]. Gonthier J considered the different standards between the law of France and of Quebec – which, generally speaking, applies the common law standard – to be significant in relation to the approaches taken by the courts of those countries:  Laferrière v Lawson [1991] 1 SCR 541 at 601-603. Professor Khoury considers that the strictness with which French courts approach the high standard of proof may have led them to resort to the loss of chance solution[39].

  147. In Germany, where there is also a high standard of proof, the problem of proof in negligence cases is dealt with by a reversal of the burden of proof. This arises in cases of medical negligence[40]. Where it is proved that a doctor was grossly negligent, the doctor must prove that his or her actions were not the cause of the injury. The term "gross negligence" is not applied restrictively. An action which is contrary to generally acknowledged rules of medical treatment, such as not performing a necessary investigation, will suffice[41].

  148. The standard of proof required by the common law already admits of some uncertainty in proof of causation. As Lord Hoffmann observed in Gregg v Scott, the wholesale adoption of possible rather than probable causation as a condition of liability is radical: [2005] UKHL 2; [2005] 2 AC 176 at 198 [90].

  149. The common law has also shown itself to be adaptable in relation to its treatment of evidence of causation in cases where there is difficulty of proof. In Snell v Farrell [1990] 2 SCR 311 at 328-329 the Supreme Court of Canada has countenanced an approach, in medical negligence cases, where inferences might more readily be drawn adverse to a defendant, because the facts lie particularly within the defendant's knowledge. The inference drawn by members of this Court in Adelaide Stevedoring Co Ltd v Forst (1940) 64 CLR 538 is noteworthy (see in particular at 563-564 per Rich ACJ and 567 per Starke J). The decisions in McGhee v National Coal Board [1972] UKHL 7; [1973] 1 WLR 1; [1972] 3 All ER 1008 and Fairchild v Glenhaven Funeral Services Ltd [2002] UKHL 22; [2003] 1 AC 32 are perhaps more controversial. In this case, however, it is not suggested that any of these approaches are possible and it is not necessary to further consider these decisions.

  150. When an issue is proved to the general standard the court treats the damage caused as certain, thus giving rise to the all-or-nothing rule of recovery. The rule is strongly criticised by those who favour acceptance of loss of chance as damage[42]. However, the rule reflects the certainty that the law considers to be necessary when attributing legal responsibility for harm caused. To replace it with a rule which limits damages awarded according to the degree of probability of causation has its own limitations. It would suggest, if not require, a degree of precision in the assessment of probabilities which is not part of the more liberal, common sense, approach presently undertaken. And, as Baroness Hale of Richmond observed in Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 233 [225], proportionate recovery cuts both ways.

  151. It would require strong policy considerations to alter the present requirement of proof of causation. None are evident. The argument that there should be compensation where breach of duty is proved simply denies proof of damage as necessary to an action in negligence. I am unpersuaded that denial of recovery in cases of this kind would fail to deter medical negligence or ensure that patients receive an appropriate standard of care. These matters appear to have been influential in Matsuyama v Birnbaum. However, a feature of that case was that the defendant was called as a witness and gave evidence that an effect of the particular contract between Mr Matsuyama's medical insurer and the doctors' practice to which the defendant belonged was that doctors had difficulty in providing patients qualifying for treatment under it with the best medical care: Matsuyama v Birnbaum 890 NE 2d 819 at 825, fn 13 (Mass 2008).

    Conclusion

  152. The appellant is unable to prove that it was probable that, had treatment by corticosteroids been undertaken earlier, the brain damage which occurred on 14 January 1991 would have been avoided. The evidence was insufficient to be persuasive. The requirement of causation is not overcome by redefining the mere possibility, that such damage as did occur might not eventuate, as a chance and then saying that it is lost when the damage actually occurs. Such a claim could only succeed if the standard of proof were lowered, which would require a fundamental change to the law of negligence. The appellant suffered dreadful injury, but the circumstances of this case do not provide a strong ground for considering such change. It would involve holding the respondent liable for damage which he almost certainly did not cause.

  153. The appeal should be dismissed with costs.


[1] See also at 234 per Baroness Hale.

[2] See the analysis of this case by Professor Stapleton, "Loss of the Chance of Cure from Cancer", (2005) 68 Modern Law Review 996.

[3] de la Giroday v Brough [1997] 6 WWR 585 at 598-601; Oliver (Guardian ad litem of) v Ellison [2001] 7 WWR 677 at 691-699.

[4] The decision is the subject of a "Note", (2009) 122 Harvard Law Review 1247.

[5] Crofter Hand Woven Harris Tweed Co Ltd v Veitch [1942] AC 435 at 442; Youyang Pty Ltd v Minter Ellison Morris Fletcher (2003) 212 CLR 484 at 508-509 [69]; Cattanach v Melchior (2003) 215 CLR 1 at 15-16 [22]-[23].

[6] Transport for London (formerly London Underground Ltd) v Spirerose Ltd (in administration) [2009] 1 WLR 1797 at 1813; [2009] 4 All ER 810 at 826-827.

[7] Higgins, The Golden Age of British Theatre (1880-1920), entry "Sir Seymour Hicks (1871-1949)", (2009); Hartnoll (ed), The Oxford Companion to the Theatre, 3rd ed (1967) at 443.

[8] "Tort Recovery for Loss of a Chance", (2001) 36 Wake Forest Law Review 605 at 617-618.

[9] See pars (d) and (e) of the passage quoted by Gummow ACJ at [35].

[10] See Prosser, Handbook of the Law of Torts, 2nd ed (1955) at 165; Glanville Williams and Hepple, Foundations of the Law of Tort, 2nd ed (1984) at 60.

[11] Williams v Morland [1824] EngR 224; (1824) 2 B & C 910 at 916 [107] ER 620 at 622], referred to in Brunsden v Humphrey (1884) 14 QBD 141 at 150 per Bowen LJ.

[12] Fitzgerald v Penn (1954) 91 CLR 268 at 277; The National Insurance Co of New Zealand Ltd v Espagne (1961) 105 CLR 569 at 590; March v Stramare (E & M H) Pty Ltd (1991) 171 CLR 506 at 515 per Mason CJ, 523 per Deane J, 531 per McHugh J.

[13] Chappel v Hart (1998) 195 CLR 232 at 256 [63]; Henville v Walker (2001) 206 CLR 459 at 491 [98]-[99]; I & L Securities Pty Ltd v HTW Valuers (Brisbane) Pty Ltd (2002) 210 CLR 109 at 128 [56].

[14] March v Stramare (E & M H) Pty Ltd (1991) 171 CLR 506 at 515 per Mason CJ, 524 per Toohey J, 531 per McHugh J.

[15] Mallett v McMonagle [1970] AC 166 at 176 per Lord Diplock; Malec v J C Hutton Pty Ltd (1990) 169 CLR 638 at 642-643 per Deane, Gaudron and McHugh JJ.

[16] Bank of NSW v The Commonwealth (1948) 76 CLR 1 at 340 per Dixon J, cited in Amaca Pty Ltd v Ellis (2010) 263 ALR 576 at 592 [70] per French CJ, Gummow, Hayne, Heydon, Crennan, Kiefel and Bell JJ.

[17] As observed by the Court of Appeal: Gett v Tabet [2009] NSWCA 76; (2009) 254 ALR 504 at 557 [258].

[18] CES v Superclinics (Australia) Pty Ltd (1995) 38 NSWLR 47; Qantas Airways Ltd v Cameron [1996] FCA 1483; (1996) 66 FCR 246; Gavalas v Singh [2001] VSCA 23; (2001) 3 VR 404; Rufo v Hosking [2004] NSWCA 391; (2004) 61 NSWLR 678; State of New South Wales v Burton (2006) Aust Torts Reports ¶81-826.

[19] Referring to Stapleton, "The Gist of Negligence – Part I: Minimum Actionable Damage", (1988) 104 Law Quarterly Review 213; Stapleton, "The Gist of Negligence – Part II: The Relationship Between 'Damage' and Causation", (1988) 104 Law Quarterly Review 389 at 392-394.

[20] And also the Netherlands, Italy, Portugal and Spain – see Winiger et al (eds), Digest of European Tort Law, Volume 1: Essential Cases on Natural Causation, (2007) at 589, and Belgium: see Graziano, "Loss of a Chance in European Private Law: 'All or Nothing' or Partial Liability in Cases of Uncertain Causation", (2008) 16 European Review of Private Law 1009 at 1027.

[21] Zweigert and Kötz, Introduction to Comparative Law, 3rd ed (1998) at 617-618.

[22] Markesinis, "The Not so Dissimilar Tort and Delict", (1977) 93 Law Quarterly Review 78 at 88, referring to Catala and Weir, " Delict and Torts: A Study in Parallel (Part III)", (1964) 38 Tulane Law Review 663 at 664.

[23] van Gerven et al (eds), Torts: Scope of Protection, (1998) at 32.

[24] Germany, Austria, Greece, Norway, Estonia and Lithuania – see Winiger et al (eds), Digest of European Tort Law, Volume 1: Essential Cases on Natural Causation, (2007) at 589.

[25] Germany: see Winiger et al (eds), Digest of European Tort Law, Volume 1: Essential Cases on Natural Causation, (2007) at 590, fn 125.

[26] Winiger et al (eds), Digest of European Tort Law, Volume 1: Essential Cases on Natural Causation, (2007) at 590-591.

[27] Matters considered significant in Matsuyama v Birnbaum 890 NE 2d 819 at 838 (Mass 2008) per Marshall CJ.

[28] Khoury, Uncertain Causation in Medical Liability, (2006) at 111.

[29] Khoury, Uncertain Causation in Medical Liability, (2006) at 113.

[30] Khoury, "Causation and Risk in the Highest Courts of Canada, England, and France", (2008) 124 Law Quarterly Review 103 at 125-126.

[31] Khoury, "Causation and Risk in the Highest Courts of Canada, England, and France", (2008) 124 Law Quarterly Review 103 at 126.

[32] The respondent did not pursue this issue on the appeal.

[33] Khoury, "Causation and Risk in the Highest Courts of Canada, England, and France", (2008) 124 Law Quarterly Review 103 at 126.

[34] Winiger et al (eds), Digest of European Tort Law, Volume 1: Essential Cases on Natural Causation, (2007) at 591.

[35] Of the Supreme Judicial Court of Massachusetts.

[36] Savatier, Une faute peut-elle engendrer la responsabilité d'un dommage sans l'avoir causé?, D 1970 Chron 123 at 124, cited in Laferrière v Lawson [1991] 1 SCR 541 at 574.

[37] As suggested by Clermont and Sherwin, "A Comparative View of Standards of Proof", (2002) 50 American Journal of Comparative Law 243 at 258.

[38] It is suggested that because civil law countries had no experience of juries, they did not develop notions of probability – see Clermont and Sherwin, "A Comparative View of Standards of Proof", (2002) 50 American Journal of Comparative Law 243 at 257.

[39] Khoury, Uncertain Causation in Medical Liability, (2006) at 137.

[40] Jansen, "The Idea of a Lost Chance", (1999) 19 Oxford Journal of Legal Studies 271 at 276-277; see also Fleming, "Probabilistic Causation in Tort Law", (1989) 68 Canadian Bar Review 661 at 670-671.

[41] Jansen, "The Idea of a Lost Chance", (1999) 19 Oxford Journal of Legal Studies 271 at 277.

[42] See Matsuyama v Birnbaum 890 NE 2d 819 at 830 (Mass 2008) per Marshall CJ; Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176 at 183 [15] per Lord Nicholls of Birkenhead.


Representations

B W Walker SC with J L A Lonergan and J Chambers (instructed by Slater & Gordon Lawyers) for the appellant.

N J Young QC with J K Kirk and K C Morgan (instructed by Blake Dawson Lawyers) for the respondent.


all rights reserved