Gett v Tabet [2009] NSWCA 76 | Case summary | Zoe

NSWNSWCA

Gett v Tabet [2009] NSWCA 76

[2009] NSWCA 76

Court of Appeal (NSW)|2009-04-09|Before: Allsop P, Beazley JA, Basten JA, Studdert J

View on NSW Caselaw

At a glance

Source facts

Court

Court of Appeal (NSW)

Decision date

2009-04-09

Before

Allsop P, Beazley JA, Basten JA, Studdert J

Source

Original judgment source is linked above.

Judgment

Source: NSW Caselaw (www.caselaw.nsw.gov.au). Converted to plain text by ZoeLaw. Not the official version.

Appeal Outcome: Appeal dismissed with costs by the High Court, 21 April 2010 [2010] HCA 12

New South Wales

Court of Appeal

CITATION: Gett v Tabet [2009] NSWCA 76

This decision has been amended. Please see the end of the judgment for a list of the amendments.

HEARING DATE(S): 30 July 2008, 31 July 2008, 1 August 2008

JUDGMENT DATE:

9 April 2009

JUDGMENT OF: Allsop P at 1; Beazley JA at 1; Basten JA at 1

DECISION: (1) Allow the appeal and set aside orders (2) and (4) made in the Common Law Division on 9 February 2007;

(2) In lieu thereof, order:

(a) judgment for the second defendant, and

(b) the plaintiff to pay the second defendant’s costs of the trial;

(3) Order the respondent to pay the appellant’s costs of the appeal;

(4) Grant the respondent a certificate under the Suitors’ Fund Act 1951 (NSW), if otherwise qualified;

(5) Order the cross-appeal be dismissed;

(6) Order the cross-appellant to pay the cross-respondent’s costs of the cross-appeal.

CATCHWORDS: APPEAL – national integrated legal system – place of intermediate appellate courts within structure of Australian judicial system – departing from existing authority – principle of restraint – certainty in the law – English and Australian authority on doctrine of precedent - APPEAL – precedent – standard of reconsideration – Court of Appeal’s power to depart from previous decisions of Court or co-ordinate courts – nature and quality of error – circumstances in which Court will depart from its earlier decisions – “plainly wrong” – “compelling reasons” – whether Rufo v Hosking should be followed – whether previous decision can be reconciled with principle or with authority – jurisprudential and pragmatic considerations - APPEAL – rehearing – nature of challenge to judgment below – appellate court’s obligation to reach own conclusion – threshold for appellate intervention –limitations on appellate court to comprehend fully evidence and process of trial – resolution of medical experts’ competing views – Supreme Court Act 1970 (NSW), s 75A - TORTS – medical negligence – harm – loss of a chance of better medical outcome – difficulties in identifying harm suffered – whether doctrine consistent with nature of harm in Civil Liability Acts – Civil Liability Act 2002 (NSW), Pt 1A, s 5 - TORTS – negligence – causation – loss of a chance of better medical outcome – whether Rufo v Hosking should be followed – whether Rufo should be followed – distinction between causation of loss and assessment of proven loss – alteration of principle of causation – proof of causation based on creation of risk, not damage on balance or probabilities – whether loss of a chance consistent with causation requirements in Civil Liability Acts – Civil Liability Act 2002 (NSW), Pt 1A, s 5E - TORTS – negligence – distinction between loss of a chance of better medical outcome and deprivation of commercial opportunity - TORTS – negligence – loss of a chance of a better medical outcome – coherence with principles of tort law – expansion and context of negligence law – review of authorities - WORDS AND PHRASES – “common law of Australia” – “compelling reasons” – “harm” – “loss of a chance” – “plainly wrong” – “rehearing”

The approach in Fox v Percy was reiterated in unequivocal terms in CSR Ltd v Della Maddalena [2006] HCA 1; 224 ALR 1; 80 ALJR 458, in particular at [13]-[22]; see also the other authorities discussed in Costa v The Public Trustee of New South Wales [2008] NSWCA 223.
In order to determine whether appellate intervention is justified in a borderline case, such as the present, it is necessary to identify with some precision the nature of the challenge to the judgment reached at trial. Procedural steps aside, it is conventional to analyse the judicial process by reference to a number of separate exercises undertaken by the trial judge. The first is to identify the nature of the claims and defences and to select the legal principles relevant to determination of those claims and defences. Secondly, having identified the relevant issues, it is necessary to assess the evidence, whether oral or documentary, relevant to each issue. To the extent that there is conflict between the evidence with respect to primary relevant facts, those conflicts must be resolved. Thirdly, in order to satisfy the relevant legal principles, it is usually necessary to draw inferences from the primary facts, or make evaluative judgments or characterisations. (These steps may be variously identified and the order is not assumed to be the same in all cases: see Azzopardi v Tasman UEB Industries Ltd (1985) 4 NSWLR 139 at 156G.)
As made clear in Fox v Percy, on an appeal by way of rehearing, the Court is not restricted to correcting errors of law. Rather, and subject to an important qualification, the Court is required to assess the evidence, inferences and evaluative judgments required to be made to determine the case. So much is treated as its statutory obligation. In some cases, the trial judge is required to exercise a discretion in granting relief. Where a decision may fall within a range, such as the assessment of damages for pain and suffering (or the imposition of a sentence in a criminal matter), the court may decline to intervene unless the trial judge has been shown to have acted erroneously in point of principle. No such question arises on the present appeal.
The qualification referred to above involves the limitations on the ability of the appellate court to comprehend fully the evidence and process of the trial. The appellate court suffers from three kinds of limitation in that regard: (a) not having seen or heard witnesses give evidence and not having seen the whole of the evidence develop and fall out at the trial;

This list is not intended to be exhaustive, but it is useful to bear in mind the points at which an appellate court will need to consider the limitations on its own ability to assess the decision from which the appeal is brought. Those limitations are not restricted to circumstances where findings of credibility have been made. Nor are those limitations restricted to the notion of “an unexpressed and unstated ‘subtle influence of demeanour’”, as described by Kirby J in CSR Ltd v Della Maddalena at [44]. In recent times the limitations of the use of the “subtle influence of demeanour” in the assessment of witnesses giving oral testimony has been pointed out: see Fox v Percy at [28]-[31]; and CSR v Della Maddalena at [180]. The limitations include a degree of wariness about the risks of intuitive judgment which has led to a greater emphasis being placed on objective circumstances, such as consistency with contemporaneous records, by appellate courts.
In the present case, the trial judge was required to resolve various conflicts between medical experts, including conflicts as to the proper emphasis which should have been given to the early history of headaches, the significance of imprecise or absent symptoms, the appropriate diagnosis in the circumstances, the appropriate treatment once a proper diagnosis was made and the likely effect of competing forms of treatment. Resolution of these conflicts required an assessment to be made of the persuasiveness of the various medical opinions. In addition, an assessment was required of the appellant himself. He, as the treating practitioner, was required to identify the material upon which he relied and explain his process of diagnosis in relation to events which had occurred some 16 years earlier. With the benefit of hindsight, and repeated reviewing of the circumstances, it becomes easier to identify key events in the medical history. It is far more difficult to put out of one’s mind the knowledge of devastating harm which befell the respondent so as to assess the conduct of the appellant against the events as they had unfolded before him. Even to identify points at which mistakes were made is only a step in the process: it is then necessary to characterise the mistakes as want of reasonable care, rather than error of judgment. The trial judge’s conclusions involved the assimilation and assessment of this large body of evidence.
In Guest v The Nominal Defendant [2006] NSWCA 77 at [102], Ipp JA commented that the reason the conclusion of the trial judge commands respect, notwithstanding the appellate obligation to reach its own conclusion, was expressed by Kirby J in Aktiebolaget Hässle v Alphapharm Pty Ltd [2002] HCA 59; 212 CLR 411 at [97]. Ipp JA said, at [102]: “In giving ‘respect and weight to the conclusion of the trial judge,’ regard must, I think, be had to the fact that her Honour saw and heard the expert witnesses. The reason for this was expressed by Kirby J in Aktiebolaget Hassle v Alphapharm Pty Ltd … where his Honour observed that it will not always possible for judges, when arriving at a factual conclusion: ‘… [t]o express all of the considerations that lead them to the evaluative and partly intuitive conclusion required in the particular case. Any exposition of judicial reasons explaining such factual findings is ‘inherently an incomplete statement of the impression which was made upon [the judge] by the primary evidence’. ( Biogen Inc v Medeva Plc (1997) RPC 1 at 45 per Lord Hoffmann).”
The comments of Lord Hoffmann in Biogen referred to by Kirby J have been influential, being followed and applied in this Court in Dobler v Halverson [2007] NSWCA 335; 70 NSWLR 151 at [52]; Williams v Minister Aboriginal Land Rights Act 1983 [2000] NSWCA 255; Aust Torts Rep ¶81-578 at [137] and in the Full Court of the Federal Court in Branir Pty Limited v Owston Nominees (No 2) Pty Limited [2001] FCA 1833; 117 FCR 424 at [26]. In Biogen Lord Hoffmann said, at [45]: “The need for appellate caution in reversing the judge's evaluation of the facts is based upon much more solid grounds than professional courtesy. It is because specific findings of fact, even by the most meticulous judge, are inherently an incomplete statement of the impression which was made upon him by the primary evidence. His expressed findings are always surrounded by a penumbra of imprecision as to emphasis, relative weight, minor qualification and nuance … of which time and language do not permit exact expression, but which may play an important part in the judge's overall evaluation. It would in my view be wrong to treat Benmax as authorising or requiring an appellate court to undertake a de novo evaluation of the facts in all cases in which no question of the credibility of witnesses is involved.”
Embedded within the approach required by all these authorities is the demonstration of error: see the discussion in Costa and Branir.
Here, the trial judge’s conclusion did not depend on demeanour or credit. Rather, there were two competing views as to the reasonableness of the appellant’s decision that a lumbar puncture be performed. The competing views, which we will discuss in detail below, depended upon a number of factors including the respondent’s presentation on 11 January 1991, the history of chickenpox, the history of headaches and the occurrence of a neurological event. The resolution of the competing views of honest, skilled witnesses involved a combined assessment of the way in which the experts gave their evidence and the rational force of their views: Guest at [105].
To succeed here, the appellant must satisfy this Court that the trial judge came to the wrong conclusion. Where no error can be identified and the conclusion of his Honour is not itself implausible, this Court should properly give weight to the view formed by the trial judge, because to do so is to acknowledge the fact that in various respects, his Honour enjoyed a beneficial position in resolving conflicting evidence, drawing inferences and making the ultimate evaluative judgment.
That said, our task is to review the record in detail in fulfilment of the requirements of the Supreme Court Act, s 75A. D. Breach of duty

There was no issue but that the appellant’s care of the respondent should be measured by reference to the standard of care required of a specialist paediatrician: Rogers v Whitaker [1992] HCA 58; 175 CLR 479, especially at 487; F v R (1983) 33 SASR 189 per King CJ at 190-191; Rosenberg v Percival [2001] HCA 18; 205 CLR 434 at 441-442 [16].
The particular issue in this case is whether the appellant breached his duty of care. This is a question of fact, albeit one involving the application of normative standards: Nagle v Rottnest Island Authority [1993] HCA 76; 177 CLR 423; and Vairy v Wyong Shire Council [2005] HCA 62; 223 CLR 422. In particular, the question “did the [alleged tortfeasor] act or fail to act reasonably?” must be resolved prospectively. As Hayne J explained in Vairy, at [105], resolving the factual question whether there was a breach of duty: “…is not to be undertaken by looking back at what has in fact happened, but by looking forward from a time before the occurrence of the injury giving rise to the claim.”
The appellant accepts that the trial judge correctly stated the applicable principles of law, but contends that his Honour gave undue weight to the opinions of other doctors and in particular, to doctors from areas of specialisation different from his own. (2) Initial history and treatment
On 14 January 1991, the respondent was diagnosed as having a brain tumour known as a medulloblastoma. That diagnosis was made as a result of a CT scan carried out on that day. The respondent’s history leading to the diagnosis of medulloblastoma was that she first came under the care of a paediatric specialist, Dr Mansour, on 28 December 1990, at the Royal Alexandra Hospital for Children, after a referral from her uncle, Dr Sheiban, a general practitioner. At that time (28 December), she had a history of persistent headache and vomiting for 10 days, that is, from about 18 December.
Dr Mansour treated the respondent on the basis that she had a streptococcal infection. He prescribed penicillin and discharged her the same day. Dr Sheiban referred her back to hospital the next day, 29 December, because the vomiting and headaches had persisted. She was re-admitted to hospital and remained there until 31 December 1990, under the care of Dr Mansour. On admission, there was no evidence of papilledema (that is, swelling of the optic nerve due to pressure). The relevance of the presence or absence of papilledema is discussed below. Various tests were performed whilst she was in hospital, but all returned a negative result and all neurological examinations up to the date of discharge on 31 December were normal. By the date of discharge, she had a chickenpox rash. The respondent’s brother had had chickenpox, his rash, or lesions, first appearing on 16 December 1990. It was accepted that the respondent had contracted chickenpox on about this date from exposure to her brother’s chickenpox.
Between the date of her discharge on 31 December 1990 and 11 January 1991, the respondent remained at home, suffering from chickenpox. Her headaches and vomiting persisted and, on 11 January, either her uncle or grandfather, also a general practitioner, referred the respondent to the appellant, who was caring for Dr Mansour’s patients while Dr Mansour was away. The appellant arranged for her re-admission to hospital that day.
The appellant provisionally diagnosed the respondent as having post-chickenpox meningitis, and he ordered that a lumbar puncture be performed that day (11 January). That attempt was unsuccessful due to the respondent’s distress in undergoing the procedure. The respondent remained in hospital.
At about 11 am on 13 January, the respondent had a neurological episode which we will describe in more detail later. The appellant was notified. He directed that the lumbar puncture be carried out urgently. This was done.
On 14 January, the respondent suffered a serious deterioration in her condition, suffering a seizure. A CT scan was performed urgently. This revealed the presence of the medulloblastoma. A right frontal intraventricular drain was inserted by Dr Maixner, a neurological Registrar at the hospital, to relieve intracranial pressure.
On 16 January 1991, Mr Johnston, a neurosurgeon, assisted by Dr Maixner, operated to remove the tumour. The removal was only partially successful, leaving an attachment in the floor of the fourth ventricle. Chemotherapy was undertaken between 26 February and 7 May 1991 and radiotherapy from 20 May to 2 July 1991. The respondent is now significantly disabled.
The central allegation of negligence against the appellant was that he should have arranged a CT scan when the respondent came under his care on 11 January 1991 and certainly before the need arose for the insertion of the intraventricular drain on 14 January 1991. It was also alleged that the lumbar puncture should not have been ordered and that it was this procedure, undertaken on 13 January 1991, that caused the deterioration in the respondent’s condition on 14 January 1991. It was alleged that had the appellant promptly arranged for a CT scan and not ordered the lumbar puncture, the respondent’s treatment would have been different, both in timing and content from that undertaken, with a better outcome for her.
The trial judge found that the respondent was suffering from a medulloblastoma that had been growing for over two years and was increasing in size during the period Dr Mansour was treating her and that Dr Mansour failed to diagnose this. His Honour also found, at [48 (ii)], that had Dr Mansour ordered a CT scan in December, it would probably have established the presence of the tumour. The trial judge found, however, that Dr Mansour was not negligent in his treatment of the respondent. In particular, his Honour held that given the virtually unanimous paediatric and neurological evidence, Dr Mansour was not negligent in failing to order a CT scan in December.
Before proceeding to consider the questions raised on appeal, it is important to understand and put into perspective a matter central to an understanding of the claim made against the appellant. That is the symptoms that accompany chickenpox. (3) Chickenpox and the prodromal period
Chickenpox is a viral illness. The accepted medical evidence was that following exposure to chickenpox, the virus develops within the body without symptoms for a period of 14-16 days (with a maximum 10-21 days), and is followed by a period of non-specific symptoms such as headache, typically for 24-48 hours, but up to 1-5 days, prior to the appearance of lesions.
There was some confusion during the hearing, as well as on the appeal, in relation to terminology and in particular, as to the meaning of the “prodromal period” and “incubation period” of chickenpox.
Dr Allen, a paediatrician, used these terms interchangeably, maintaining that “the incubation period is exactly the same as the prodromal” period. Dr Allen stated in his expert report that the prodromal period of varicella is 14-16 days, although it could range from 10-21 days. Dr Allen explained in examination in chief that the prodromal period is the time from contact with the varicella vesicles (the blister-like lesions) until the beginning of symptoms. He then described a prodromal symptom period, before the appearance of the rash, during which the patient feels unwell, has fever, headache, anorexia, and the signs of a viral infection. Muscle pain and sore eyes might also be a feature. He said that the prodromal symptoms may last for up to a week before the lesions present, but usually only last 3-5 days. The lesions, once they appear, may last up to ten days. Notwithstanding Dr Allen’s initial elision in terminology between the incubation period and the prodromal period, he agreed in cross-examination that headaches and vomiting do not occur during the incubation period.
Dr Knight, a paediatric intensive care physician, drew a distinction between the incubation period in the prodromal period, describing the incubation period of varicella as the time from initial exposure to a person whose chickenpox lesions had appeared, to the onset of symptoms. He said that this period is usually 14-16 days after exposure, but could range from 10-21 days. Dr Knight described the “prodrome” as a set of non-specific symptoms that predate the typical features of an illness and the “prodromal period” as the period in which those symptoms appeared. In the case of varicella, the prodromal period was 24-48 hours and the prodromal symptoms typically included fever, malaise and headache. In cross-examination, Dr Knight accepted the prodromal symptom period could extend for up to five days at the most. Beyond that period, he did not consider the respondent’s symptoms could be attributable to the prodrome of chickenpox.
The appellant agreed that the prodromal period for chickenpox is usually short, could be variable between patients and could be longer than 24 hours.
Dr Mansour’s evidence was that symptoms could occur during the incubation period. This evidence was inconsistent with the other medical evidence and was rejected by the trial judge.
We will adopt the terminology, used by Dr Knight, of incubation period and prodromal period, noting that the trial judge also used this terminology. However, the important question in this case, as will be apparent from the section from [77] below entitled “The headaches and the timing issue”, is the period in which symptoms such as headache and malaise were experienced by the respondent. (4) History of the respondent’s care under the appellant
In the period prior to her re-presentation to hospital on 11 January 1991, the respondent had severe chickenpox lesions all over her body and face, as well as on her tongue. She was lethargic and was losing weight, and during this time she vomited on at least two occasions. She complained of stomach pain and headaches, although the significant complaint was of the rash and chickenpox lesions on her body.
On 10 January 1991, the respondent woke “screaming from headaches” and became less responsive to her parents. On 11 January, the respondent’s father described her as being “the same or a bit worse than the day before”. She was inactive, pale, tired, complaining of pain. He thought she vomited, but “she was really indicating that she has pain in her head or headache”.
The evidence of the respondent’s mother was to the same effect, although she said what she most noticed during this period was that the respondent was getting very pale and losing a lot of weight. She was very tired and not interacting.
When the chickenpox rash settled down, the respondent’s parents observed that she was still vomiting and still had headache. The respondent’s mother contacted either her father or brother (both general practitioners), who referred her to the appellant.
The appellant saw the respondent in his consulting rooms at the hospital on 11 January 1991. The appellant observed that the respondent was dehydrated, irritable, listless and did not interact well. Although he did not have notes of the consultation, the appellant remembered that the respondent’s parents said that she wasn’t drinking much and that she was complaining of headaches. The appellant conducted a full neurological examination, including examining her cranial nerves, her power, tone and reflexes and the ocular fundi to see if there was any evidence of raised intracranial pressure. In his evidence, the appellant said that the focus of his attention when he first saw the respondent was her deterioration in the preceding few days, that is, “her immediate problems”, including her headaches and general condition.
As the respondent had just had chickenpox, the appellant was concerned about the possibility of inflammation of the lining of the brain, that is, of the meninges, and the possible inflammation of the underlying brain itself. The inflammation of the lining of the brain is known as meningitis. Where the lining of the brain is swollen and inflamed, there is often an underlying inflammation of the brain tissue itself. That condition is known as encephalitis. The diagnosis of encephalitis is made by an examination of the spinal fluid, which is achieved by undertaking a lumbar puncture, involving the insertion of a needle into the base of the spine to obtain a sample of the fluid surrounding the brain. The appellant said in his evidence that if infection and inflammation of the brain were suspected, then a lumbar puncture was the first test that should be undertaken.
The appellant arranged for the respondent to be admitted to hospital under the care of Dr Kohn. The appellant wrote a note to Dr Kohn in the following terms: “Thank you for admitting [the respondent] who has clinical signs of meningitis. She has resolving varicella rash. She needs L.P. and IVI fluids.”

Dr Kohn examined the respondent. Her pupils were found to be normal. He examined various cranial nerves and tested for neck stiffness. There were no signs of neurological abnormality. The respondent was also examined by Dr Blissta on the same date. According to the hospital notes (apparently of Dr Blissta) there were no meningeal signs and the respondent was afebrile, conscious and well-oriented.
The hospital notes record that Dr Kohn unsuccessfully attempted a lumbar puncture on 11 January and discussed this result with the appellant. The attempt had been abandoned because of the respondent’s distress. From 3.30 pm until 6.30 am on 12 January, the respondent’s neurological condition was stable, although the nursing night report recorded that the respondent had had an unsettled night and had vomited three times. Panadol had been given at 12.40 am with minimal effect. The respondent’s neurological status remained satisfactory for the rest of that day, although the nursing notes record that the respondent was unsettled and irritable at times. She was afebrile and was not tolerating clear fluids. She vomited twice, probably at about 9.45 am, when it appears that attempts were made to provide her with fluids.
The appellant examined the respondent in hospital on 12 January 1991, after first examining Dr Kohn’s notes. The appellant’s notes record that she was still vomiting and was afebrile, and he diagnosed a viral infection. The clinical picture at this time was that the appellant was still listless and irritable. However, there were no focal signs (that is, there was no weakness of one side compared to the other and her pupils were equally reactive). The appellant stated in his evidence that the respondent’s clinical condition was still consistent with meningoencephalitis. The appellant decided to defer a lumbar puncture for at least 24 hours because the respondent was “clearly irritable and crying”.
During the course of the evening of 12 January, the respondent complained of headache at about 11 pm and was given Panadol, with good effect. She slept until 4 am, when she woke, feeling nauseous and she vomited a small quantity of yellow-green fluid. She was again given Panadol with good effect. She was afebrile and neurological observations remained within normal limits.
On 13 January 1991, at 11 am, the respondent’s father observed that the respondent was lying still and staring into space. He called to the nurse for her attention. The nursing note records that although the father called out to nursing staff because the respondent was “staring and unresponsive”, she was observed to be “responsive but irritable and drowsy and complaining of headache”. However, and importantly, her pupils were noted by the nurse to be unequal and her right pupil was not reactive.
Dr Jones, the registrar on duty, was called. His hospital note states “Viral encephalitis” and “Now [increasing] headache, drowsy … pupils equal and [reactive]” (emphasis added). He also noted that the respondent was responsive to commands. Dr Jones discussed the respondent’s condition with the appellant, including the observations of the nurse. The appellant examined the respondent (although he was unable to remember the details) and ordered the lumbar puncture.
The appellant said he was still concerned about the respondent’s condition and the reason he ordered the lumbar puncture was that he still did not have a definitive diagnosis. This decision was supported in the evidence by Mr Johnston, Dr Allen and Professor Watson, neurologist.
The lumbar puncture was performed at approximately 11.30 am on 13 January. The pathology results of the lumbar puncture were not obtained until 29 January, although the delay was undoubtedly because the events of 14 January overtook all other matters. The nursing notes for the balance of the day record that the respondent’s condition was stable and that her pupils were “equal and reacting”.
The respondent’s father said that there was another incident at about 9 pm on 13 January, when the respondent was again staring and he called the nurse. He also said a doctor attended. There are no hospital notes of this incident. The notes do record that the respondent had an unsettled night. She was seen by Dr Lane, a resident medical officer, at 1 am on 14 January, who noted that persisting headaches were not relieved by regular doses of paracetamol and codeine. Dr Lane recorded that the respondent’s neurological signs were stable. He did observe that there was “neck guarding and stiffness”.
At midday on 14 January, the respondent had a seizure, preceded by an episode, about 15 minutes earlier, of staring into space. She was unresponsive on examination. Her pupils had deviated to the left side. Nystagmus (rapid, involuntary eye movement) was observed and the respondent’s pupils were of different dimensions. The seizure lasted for approximately 15 minutes. Dr Ouvrier, neurologist, was called and a CT scan was undertaken at 12.40 pm, which detected the medulloblastoma. The CT scan report was of a posterior fossa tumour with obstructive hydrocephalus.
Following the scan, the respondent was referred for an EEG. The EEG report, which was signed by Dr Ouvrier, concluded: “A very abnormal recording with a marked generalized excess of slow wave activity. Episodes of tonic posturing were unassociated with epileptogenic activity. The appearance is consistent with a severe elevation of intracranial pressure and/or varicella encephalitis .” (Emphasis added)
The appellant’s condition continued to deteriorate and shortly after 2.15 pm a ventricular drain (CFS drainage) was inserted by Dr Maixner.
On 15 January, Dr Maixner observed the respondent had improved with the CFS drainage. She was able to breath spontaneously, her pupils were equal and reactive, her eyes were midline, her corneals were present, although better on the left, her left arm flexed to pain and she crossed her legs spontaneously.
On 16 January, Mr Johnston carried out an operative procedure to remove the tumour. Dr Maixner assisted. Dr Maixner recorded that: “… a large tumour was found arising from the wall of the fourth ventricle on the right side. The tumour had spread in a subpial fashion over the surface of both cerebellar hemispheres and was protruding over the foramen of Magendie. I documented that an incomplete removal of the tumour was achieved.”
Following surgery, the respondent had sessions of chemotherapy between February and May 1991. Following that, she had radiotherapy between May and July 1991. (5) Allegations of negligence against the appellant
The trial judge recorded three essential allegations of negligence made against the appellant, as follows: “(i) The appellant was negligent in not arranging for a CT scan;

The trial judge found, at [122], that there was support in the medical evidence for the reasonableness of the appellant’s provisional diagnosis of post-chickenpox encephalitis made on 11 January. In this regard, it should be noted that the respondent saw two other practitioners at the hospital on 11 January: Dr Kohn and Dr Blissta, both of whom undertook a neurological examination of the respondent. No neurological signs were evident on those examinations. At this stage, the only significant indication which was inconsistent with the provisional diagnosis made by the appellant was that the plaintiff’s headaches and vomiting predated the chickenpox prodromal period and were therefore unlikely to be associated with post-chickenpox meningitis or encephalitis.
It is not clear whether the appellant appreciated at the time he decided to proceed with the lumbar puncture on 13 January, that the respondent had a history of headaches going back to 18 December 1990. His Honour found that if the appellant did not in fact know that history, the means of becoming aware of it were available, both from the hospital records relating to the first admission and from the respondent’s parents, one of whom was always at the respondent’s bedside. This finding is of particular significance, as it was made in the context that his Honour found that the reason the appellant had arranged for the respondent’s admission to hospital on 11 January was to explore her history.
His Honour returned to this issue at [192], where he made the following finding: “… the provisional diagnosis upon which the appellant was proceeding on 13 January 1991 did not satisfactorily account for the history of headaches and vomiting from 18 December 1990 onwards. In evidence to which I earlier referred (at [138]) the appellant said that exploring the history of headaches was "a reason for admission", but as at 13 January 1991 there was no satisfactory explanation for their duration, and the episode at about 11.00 am on that date needed to be assessed with this consideration very much in mind.”
The appellant specifically challenges the findings in this paragraph. This challenge is central to the appellant’s attack to his Honour’s finding of breach of duty.
In finding that the appellant had breached his duty of care, the trial judge accepted the opinions of Dr Knight, Dr Hopkins, Mr Wallace, Mr Klug and Dr Maixner over the evidence of Mr Johnston, Dr Allen and Professor Watson. In particular, his Honour was persuaded by the respondent’s experts that there was an urgent need for a CT scan following the incident at 11 am on 13 January, when the respondent’s pupils were observed to be unequal and the right pupil to be not reactive.
The relevance of that incident is that it was the first recording of a neurological incident. His Honour considered that its occurrence introduced an additional element that had to be weighed by the appellant in the exercise of reasonable care. His Honour found that once that incident occurred, the appellant ought to have arranged for an immediate CT scan and ought to have waited on its results, before proceeding to have a lumbar puncture carried out. His Honour further found that had a CT scan been performed on 13 January, the medulloblastoma would have been detected and the lumbar puncture would not have been performed. In making this finding, his Honour stated that he was conscious of the difficulties that the case presented for the appellant, by reason of the concurrence of the severe infection, that is, the chickenpox and the presence of the medulloblastoma.
His Honour found no earlier negligent act or omission.
The appellant contends that although his Honour correctly stated the legal principles in determining whether there had been a breach of the appellant’s duty of care to the respondent, he arrived at the wrong answer, because he gave undue weight to expert evidence that was given with the benefit of hindsight by persons who were not specialist paediatricians.
The appellant also complains that the trial judge gave undue weight to the respondent’s history of headaches. He says that the headaches from which the appellant was suffering on 11 January were consistent with his provisional diagnosis, of post-chickenpox encephalitis. It was that provisional diagnosis that the appellant sought to test by deciding that a lumbar puncture be performed before a CT scan, should a CT scan become necessary at all. The appellant contends that as his Honour found that his provisional diagnosis of post-chickenpox encephalitis was “understandable”, his Honour thereby implicitly found that it was reasonable. Accordingly, the appellant’s decision to test for that provisional diagnosis was also reasonable and that in pursuing that course he did not fail to exercise reasonable care.
The appellant further complains that in reaching the conclusion that the appellant was negligent in failing to order a CT scan on 13 January 1991, his Honour gave insufficient weight to two important factors: first, the lack of neurological signs prior to 13 January; and, secondly, the transient nature of the neurological symptoms that occurred at 11 am on that day. (7) The headaches and the timing issue
The consideration that dominated the evidence of those experts whom his Honour accepted, was the history of the respondent’s headaches. If chickenpox-related headaches only occurred during the prodromal phase, it was necessary to determine when that period commenced. Two dates were relevant. The first was 16 December, when the respondent’s brother’s chickenpox lesions appeared. The second was 18 December, which was the date his Honour found the respondent first began experiencing headaches.
The timing was important, because, if the headaches could not be explained by chickenpox in the prodromal phase, that would have put the appellant’s provisional diagnosis of chickenpox meningitis or encephalitis into doubt so that when the neurological episode occurred on 13 January, the appellant should have sought to investigate the respondent’s neurological state by arranging for a CT scan, rather than by seeking to confirm or negate meningitis and encephalitis by way of a lumbar puncture.
The recorded history relating to headaches commenced relevantly with the consultation with Dr Sheiban and his referral to Dr Mansour. Dr Sheiban’s consultation notes of 26 December contained a history of headaches. Dr Sheiban’s referral note to Dr Mansour of 28 December 1990 was in the following terms: "Thank you for seeing this little girl, who happens to be my sister's daughter. She has been [complaining of] headache and vomiting for [10 days]. No diarrhoea. Small fever Previously healthy [Previous history]: [Urinary Tract Infection] x 2 Thank you again for this quick arrangement."
Dr Mansour’s notes taken on 28 December recorded: "O/E [on examination] Not distressed BP 100/ HS No neck stiffness Fundi Cranials No bruits Abdo soft BS [bowel sounds] N [normal] No tenderness Chest No lymphadenopathy Throat - red - spots on tonsils. Ears ok IMP [impression]? possibly URTI [upper respiratory tract infection] ? UTI [urinary tract infection] For msu [micro urine] c&s [culture and sensitivity] Then course of Pen V [penicillin] 125 mg tds [three times a day] 10 days"
On the second referral, on 29 December, Dr Sheiban’s note again referred to the history of headaches, in the following terms: "Thank you for seeing this girl, who presents with [more than 10 days] of persistent vomiting and headaches. No diarrhoea. Temp 36.7. ? Small fever early on. No signs of meningism. [Abdominal] pain, but no organmyopathy or [Lymph Nodes] enlargement. [Previous history]: [Urinary Tract Infection] x 2 Yesterday she was commenced on penicillin by Dr Mansour for streptococcal infection. Still vomiting & [complaining of] abdominal pain & headache. ? Needs I.V. fluids & [blood test] - Nil by mouth."
On 29 December, the respondent underwent two medical examinations. Dr Durman, who conducted the first examination, recorded “intermittent vomiting, headaches and abdominal pain for ten days”. A neurological examination was normal. Dr Williams conducted the second examination and it was following that examination that the respondent was admitted to hospital. It is not clear from the evidence whether Dr Durman and Dr William’s notes formed part of the hospital records.
The nursing notes (which were part of the hospital records) for 29 and 30 December record the following: “… headache and vomiting and has been unwell for 10 days. Afebrile in no apparent distress. Looks unwell and lethargic.”

The respondent’s counsel suggested to the appellant that the history of headaches and vomiting “for many weeks” was inconsistent with a varicella infection. The appellant disagreed, stating that it “is still consistent with a viral papillary, secondary to varicella”. He added: “The acute phase is over in generally five to seven days. You can often have a longer phase after that and have cerebella signs and inflammation, unusual, but it can occur.”
Dr Allen’s evidence supported the appellant’s position.
Dr Allen was of the opinion that because the neurological event at 11 am on 13 January was transient and the respondent had returned to her previous condition, the likely primary diagnosis would remain as varicella encephalitis. He pointed out that the chances of having chickenpox encephalitis and a brain stem tumour at the same time was “a very uncommon process”. As he explained: “That, to have two diagnostic processes happening at the identical time when you have a confirmed diagnosis of chickenpox which has well recognised neurologically [sic] by complications related to that. To assume that there is some other process happening at the same time is very unlikely and I, as a clinician, would have said to myself, we’ve had this episode, we are now neurologically normal. It is reasonable to proceed to do the lumbar puncture to attempt to confirm the primary diagnosis.”
Dr Allen was of the opinion that the respondent’s exposure to her brother’s chickenpox on 16 December and the history of “frontal headaches significant from 23 December 1990” (being the history Dr Allen understood had been given to Dr Mansour), was still very consistent with the respondent having an intercurrent viral illness and consistent, perhaps, with it developing into an encephalitic process or infective process.
It was pointed out to Dr Allen that the history given to Dr Mansour was not in the terms Dr Allen had understood. Rather, the position was that from 23 December, “the frontal headaches had a severity which was significant”. Dr Allen stated that he still considered that this indicated an infective process and that was how he would have considered it. He said he would monitor the course of the headache and the illness for a period of time before he would do other investigations, because this was a common presentation for viral infection in children in hospital. Dr Allen considered that the respondent’s history at the time of the first admission to hospital of sleeping well at night was significant, because that would reassure him, as a clinician, that there was no evidence of raised intracranial pressure. By contrast, where there was a history of headache waking a child at night, that would be a matter for concern.
Dr Allen accepted in cross-examination that the respondent’s history of headaches starting on 18 December, taken with the history that the chickenpox lesions appeared on 31 December, meant that the headaches could not be related to chickenpox. Dr Allen said that the headaches may have been from some other viral infection. He said it was common to see people with a prolonged fever which related to two infections. Dr Allen expressed the opinion that it was easy in retrospect to be critical of the provisional diagnosis made by the respondent. Importantly, he said it was not bad clinical medicine for such a diagnosis to have been made.
Mr Johnston was of the opinion that as at 11 January 1990, there were not sufficiently specific physical signs to warrant a CT scan: “Q. …You said, in the report of 24 January 2001, that you would not have personally ordered a CT scan at that time, meaning 11 January 1991, when the child was admitted back under [the appellant] with an unquestioned viral illness without evidence of intracranial hypertension. I put it to you that there was evidence of intracranial hypertension by that time, namely persistent headache and vomiting, from 18 December 1990 A. By evidence I mean any physical signs.

Dr Knight, at the time he gave evidence, was a paediatric intensive care physician at the Princess Margaret Hospital for Children in Perth. Prior to that, from 1993 until approximately 2002, Dr Knight undertook “a significant amount of general paediatrics”. He said there was not a clear time when he stopped doing general paediatrics, but the amount he did gradually decreased as his work in the full-time intensive care unit increased.
Dr Knight, in his expert report dated 22 April 2002, expressed the opinion that a diagnosis of the symptoms of 10 days of headache, vomiting and abdominal pain prior to the admission on 29 December 1990, in circumstances where there was an absence of abnormal clinical signs, did not sufficiently suggest intracranial lesion to say that a CT scan should have been done prior to the respondent’s discharge on 31 December. However, he stated that at the time of admission on 11 January, the respondent’s symptoms and in particular, the duration of her headaches, were not consistent with the diagnosis of varicella encephalitis.
In his evidence in chief, Dr Knight explained that the reason he had stated that the symptoms of 10 days of headache, vomiting and abdominal pain were not explained by chickenpox, was because such symptoms are known to occur in the period up to 48 hours before the rash appears. He said that symptoms that pre-date the rash by 10 days are not explained by varicella. In his view, a differential diagnosis at that time should have included an intracranial lesion. He said that an intracranial lesion is a “reasonably common cause of headache and vomiting” in the respondent’s then age group of six years of age.
Dr Knight provided a further report, dated 13 June 2006. In addition to the hospital notes, which he had available at the time of his report dated 22 April 2002, Dr Knight also had reports from Drs Allen and Williams; a statement of assumptions from the appellant; and a report of Dr Sheiban (the appellant’s uncle). In this report, Dr Knight stated there was little support, on the material he had, for a diagnosis of meningitis on 11 January and there were no clinical signs suggestive of that condition. He said the long history of headaches and the absence of signs, should have led to a consideration of alternative causes of the respondent’s symptoms and, in particular, to the presence of a tumour.
In a report dated 13 July 2006, Dr Knight commented that a diagnosis of, inter alia, acute varicella encephalitis did not account for the long history of headache and vomiting.
In his examination in chief, Dr Knight explained that the basis for his reasoning that the symptoms of 10 days of headache, vomiting and abdominal pain at the time the respondent was admitted to hospital on 28 December 1990 were not explained by the varicella, was that the symptoms preceding the onset of the varicella rash are known to appear up to 48 hours before the rash. Symptoms pre-dating the rash by 10 days were not, therefore, explained by varicella.
It was central to Dr Knight’s evidence that the history of headache and vomiting from 18 December was critical and there had been no explanation for it. From his viewpoint, the varicella did not explain it. Dr Knight was of the opinion that at the time of the respondent’s discharge from hospital on 31 December 1990, the history of headache and vomiting had not been adequately explained. He agreed, nonetheless, that even with that history, it was appropriate at that time to adopt a conservative approach. He said it was less acceptable to do so on the second admission, because the symptoms had continued for so much longer. He was firm in his view, however, that in a case with a history of unexplained symptoms (that is, symptoms for which no definite diagnosis had been made) which might be attributable to intracranial pressure, a lumbar puncture should not be performed.
Dr Hopkins considered that a CT scan should have been performed on 11 January, because the continuing symptoms of headache and vomiting indicated raised intracranial pressure, including possibly due to a tumour.
Mr Wallace, neurosurgeon, considered it understandable that the appellant arrived at a diagnosis of post-chickenpox encephalitis, but said that the headaches and vomiting were “a red light”. Mr Wallace said that the headache and vomiting within two days of exposure to chickenpox meant that the respondent could not have chickenpox encephalitis. Mr Wallace said that the fact the headaches went on for two weeks was another “red light that there may be something else to it”, but agreed that the appellant’s provisional diagnosis was understandable. It was suggested to him in cross-examination that, in that circumstance, it would have been appropriate to consider a lumbar puncture to confirm or exclude that diagnosis. Mr Wallace agreed, but added that the problem of the headache and vomiting was still there. That was the “red light” from his point of view. He said, however, that if the headache and vomiting were put to one side, he otherwise would agree that a lumbar puncture was appropriate.
The relevance of the headaches was also a central aspect of Mr Wallace’s evidence. He provided a report dated 14 December 2005. It is relevant to set out his opinion on this aspect and, in particular, his opinion as to the appropriateness of undertaking a lumbar puncture before a CT scan. This is contained in the second and third paragraphs of his report, where he said: “… It is to be noted that this child presented with headache and vomiting in late 1990, having contracted chickenpox on 15.12.1990, and being referred to Dr Mansour who admitted her to hospital and told the parents if symptoms did not settle, she would need a CT scan. She subsequently developed a typical rash on 31.12.1990, confirming the doctor’s suspicion that the child was developing chickenpox. With continuing headache and vomiting and with a severe chickenpox rash, a review was undertaken in Dr Mansour’s rooms by his locum, [the appellant]. It is utterly understandable that with confirmed chickenpox, the diagnosis of chickenpox encephalitis or meningitis was undertaken, and it was therefore logical to consider doing a lumbar puncture to confirm the diagnosis of chickenpox encephalitis or meningitis.

Mr Wallace was cross-examined by senior counsel for the appellant on this evidence. In particular, he was asked to confirm that in the second paragraph, there was no mention that the diagnosis was inappropriate, bearing in mind the appellant’s history of headache and vomiting. He agreed that was so in respect of that evidence, but stated that the whole tenor of his report was “the concern that there is raised intracranial pressure here”. Mr Wallace accepted that a lumbar puncture to confirm a diagnosis of chickenpox encephalitis was logical. However, that did not explain the headaches and vomiting which preceded the chickenpox. He said: “So if there was only the issue of chickenpox and headache and vomiting then it is very reasonable to consider doing a lumbar puncture but with headache and vomiting it should be a red light to a clinician, an experienced clinician, that – particularly when it’s gone on for weeks.”
Mr Wallace was cross-examined as to how he could maintain that opinion, given that he had agreed the appellant’s provisional diagnosis of post-chickenpox meningitis was “utterly understandable”. Mr Wallace said: “No, I understand his processes, but … he has missed the point of the headache and the vomiting which are the key here.”
The appellant’s counsel suggested that if he was of the opinion that the appellant had “missed the point”, he should have qualified his report to state that (this was a reference to the last sentence of the second para set out above). Again, it is useful to record Mr Wallace’s answer. He stated: “The word in there is ‘to consider’. That’s where I’m couching the fact that I’m not condoning it. I’m saying I understand it should be considered. But there’s an inference in that which is perhaps too subtle or should have been more blatant, but I was really putting forward the view that the doctor’s thinking is understandable, but I didn’t go to say that I believed it was safe to do a lumbar puncture. I said it was logical to consider it and I don’t believe it was safe to do it .” (Emphasis added)
Upon further questioning, Mr Wallace became more firm. He said: “I’ve only gone as far as to say I can understand the logic of it, to consider doing a lumbar puncture. But whenever you do a lumbar puncture, what I’m implying is there are other considerations and I didn’t say ever, that I condoned the doing of the lumbar puncture in this situation, because I don’t. I don’t believe it was a safe thing to do.”
He added that when regard was had to the next paragraph of his report, then: “… it follows on that when I said, ‘I think it’s logical to consider doing a lumbar puncture’, but in this instance the danger is there. And therefore, I think what I’m trying to say however badly in the report, and however kindly as I can, was that in this particular case, it wasn’t a good idea. It was dangerous.”
Mr Wallace’s cross-examination proceeded: “Q. The role of a lumbar puncture in the scenario you were addressing in the [report] was a form of investigation after a diagnosis had been reached, provisional or otherwise. Correct? A. Yes.

It was never suggested in the case that there was evidence of either of these unequivocal signs of raised intracranial pressure on 13 January.
Dr Knight was of the opinion that a CT scan should have been done on 13 January. In his report dated 22 April 2002, Dr Knight said that the episode on 13 January, of pupil inequality, elevated blood pressure and subsequent drowsiness, was a further indicator of intracranial pathology and increased intracranial pressure. He said that drowsiness and pupil inequality were “suggestive of markedly raised intracranial pressure” and that a lumbar puncture should not have been performed when those clinical signs were present. He stated: “The suspected underlying disease is not relevant to the issue of whether a lumbar puncture should be performed in these circumstances. Immediate management of the raised intracranial pressure should be the first priority followed by a timely CT scan.”
So far as the incident on 13 January was concerned, Dr Knight considered that, not only were the unequal pupils of relevance, so was the respondent’s elevated blood pressure. Dr Knight explained that there were a number of causes of unequal pupils, but the most concerning cause was an effect on the third cranial nerve, which can be caused by pressure on that nerve due to raised pressure inside the head, that is, raised intracranial pressure.
Dr Knight stated that when there is a suspicion of raised intracranial pressure, a CT scan is, and was in 1991, used as a means of reassuring clinicians that it was reasonable to proceed with a lumbar puncture. In his opinion, a lumbar puncture on 13 January was contraindicated because of the episode of dilated pupils, which he said was a “clinically significant observation”. The episode of dilated pupils should have been combined with the history of headache and vomiting to raise even further the concern that the respondent had intracranial pathology. He said it was at that time that a CT scan was indicated and that it should have been done as soon as possible, by which he meant within the following few hours. He reiterated that the CT scan should have been done before the lumbar puncture was undertaken.
Dr Knight also disagreed it would be reasonable to proceed with a lumbar puncture after neurological signs had disappeared. His opinion in this regard was expressed in response to cross-examination that neurological signs may appear and disappear because of the wave-like nature of the course of hydrocephalus, secondary to a posterior fossa tumour. Dr Knight agreed that such pressure waves characteristically last for 10, 15 or 20 minutes and then spontaneously resolve, in the sense that the internal cranial pressure would reduce. Dr Knight still considered that in that situation, it was not safe to perform a lumbar puncture. It was suggested to Dr Knight that, in circumstances where the intracranial pressure had peaked and then reduced to a point where it was no longer causing signs, it would be safe to undertake a lumbar puncture. He said that once the signs were present then, whether or not they had resolved, it was a sufficient indication that the intracranial pressure was high and would remain high.
Dr Knight accepted that one reasonable course following the incident on 13 January was to have involved a neurologist. He said that if that had been done, the neurologist would have needed to have taken the same course, that is, arranged for a CT scan to be undertaken.
It was Dr Knight’s opinion that there was no clinical scenario where an episode of intracranial pressure could be so high as to cause focal neurological signs and then resolve to the point that the risk of it happening again, at any particular moment, would not be high. He said that a lumbar puncture was contraindicated in the presence of suspected raised intracranial pressure. He did not agree that paediatric clinicians, acting reasonably at the time of the time of the respondent’s illness, might take a different view in that regard. In particular, he did not agree that it was reasonable to continue to monitor the respondent by observation after the focal neurological signs had appeared during the incident on 13 January. It was his opinion that with an episode of abnormal neurological signs on top of the respondent’s history, something needed to be done immediately after the incident on 13 January. That “something” was a CT scan.
Mr Klug considered that the incident on 13 January could be associated with certain situations of raised intracranial pressure. However, he said it was not a clear indication, as there are other causes as to why pupils may be unequal. He said “it would raise one’s suspicion but it wouldn’t confirm [intracranial pressure]”. He considered, however, that it was a very strong indication to undertake a CT scan before even considering the use of a lumbar puncture as a diagnostic procedure. Mr Klug said that the coincidence of other symptoms during this incident would reinforce his opinion. The question asked and his answer on this were in these terms: “”Q. Doctor, if it were the fact that the parent, in reporting that incident, reported that the child was staring and unresponsive and the staff, when they observed the child after it had been reported to them, reported that the child was responsive but irritable and drowsy and complaining of headaches and the pupils were noted to be unequal and the right pupil not reactive although the other observations were within normal limits; the blood pressure was 131/85, would that in any way alter or reinforce the opinion you’ve just expressed to the need to undertake a CT scan? A. I think it would reinforce the indication to undertake a CT scan.”
Later, Mr Klug was asked: “Q. When you gave the evidence about the signs not being a clear indicator of raised intracranial pressure you were acknowledging the fact were you not that there were other possibilities, that is, other than raised intracranial pressure which may have produced the signs which were reported and observed?”

At this point it is important to appreciate the significance of papilledema, which is the swelling of the optic nerves due to raised intracranial pressure. Mr Johnston said that headache, nausea and vomiting were very non-specific symptoms, common to a wide variety of conditions. He said that if a patient was to complain, relevantly, of headache, and it was persistent, then some sort of investigation would be appropriate, depending on the circumstances. One of the tests that would be undertaken would be an examination to determine whether or not there was intracranial pressure. He said the key aspect of the examination was to look for evidence of papilledema by examining the eyes with an ophthalmoscope. Mr Johnston explained the process when examining for papilledema. He said that by using the ophthalmoscope it was possible to see the point of entry of the optic nerve into the retina. He said it was more or less in the centre of the retina and appeared as a round, sharply demarcated structure. He said with swelling of the optic nerve, the sharp margin of the nerve disappears, and the nerve head becomes raised and indistinct. He said that in later stages, haemorrhages and exudates appear in the retina.
If there was raised intracranial pressure, then there may be signs of papilledema, but not necessarily so. As Mr Johnston explained: “It’s a key sign and it may be an early sign but it may not appear and the pressure may still be quite raised. So it’s not an infallible sign and it’s not necessarily an early sign but it’s an unequivocal sign when it’s present.”
The other sign available from examination was a “sixth nerve palsy”, which Mr Johnston described as a “squint paralysis of ocular movements”. He said papilledema and sixth nerve palsy were the only real signs of pure intracranial pressure that have any degree of frequency.
Mr Johnston said that the absence of those signs would “sharply reduce” the suspicion of raised intracranial pressure. He added, “substantially reduce [sic], but not absent”. Mr Johnston emphasised the crucial importance of the lack of papilledema in this case, particularly in relation to the carrying out of a lumbar puncture. He agreed, however, that the absence of papilledema would not deter a clinician from performing a CT scan if there were other sufficient symptoms present. Such symptoms would include a history of 10 days of headache and vomiting. In this regard, Mr Johnston stressed it was important that there was an absence of any other explanation.
Mr Johnston accepted that there were some signs of raised intracranial pressure present on 13 January, but he said they were not unequivocal signs. For example, neck guarding, or neck stiffness, could be a manifestation of raised intracranial pressure, but it was by no means a specific indication.
Dr Maixner disagreed that, immediately following the incident on 13 January at 11 am, when no neurological signs were elicited on examination, the respondent would not have been regarded as an urgent patient for the undertaking of a CT scan. Dr Maixner considered that notwithstanding that it appeared the neurological symptoms had passed, the case remained urgent, given the respondent’s history of headache and vomiting. She said the absence of papilledema or sixth nerve palsy did not exclude intracranial pressure and that it was necessary to rely on the clinical history and the conscious state of the patient. From her point of view, these two signs were not essential and were not predominantly present in all patients. In this regard, her evidence accorded with that of Mr Johnston.
Dr Maixner considered that even in a child that had post-chickenpox encephalitis, the pupillary dilation and the headache and vomiting, would still give rise to a concern that the child had intracranial hypertension. She was adamant that if she had been notified on Sunday 13 January in relation to the respondent’s condition, she would have recommended a CT scan on that day. She accepted that there were differential diagnoses relating to the respondent. However, in circumstances where intracranial hypertension could not be ruled out and the other diagnoses had also not been ruled out, then, as the neurosurgery registrar, she considered it would have been her job to rule out intracranial pressure and she would have done that by having requested a CT scan on 13 January.
Dr Allen was of the opinion that the absence of papilledema would suggest to him that there was no evidence of raised intracranial pressure. He agreed that papilledema was not always found in cases of raised intracranial pressure, but was usually present, probably more than in 95 per cent of cases. He said, however, that if there was an acute extradural haemorrhage, then it would take some time, that is, hours, for papilledema to develop within the eye. Dr Allen agreed that there could be raised intracranial pressure without papilledema, but that was very infrequent, unless the intracranial pressure was very acute. Dr Allen said that in the case of a transient neurological event, where there was no evidence of papilledema, it was safe to do a lumbar puncture.
Dr Knight also disagreed that the absence of papilledema after the incident at 11 am on 13 January was an indication that whatever raised intracranial pressure in the respondent had been present, it had sharply reduced. His view was recorded in the following question and answer: “Q. Is it your evidence, so we have it clearly, that against the background of the observed focal neurological signs recorded at 11 o’clock it was not possible to perform any examination which would have given sufficient satisfaction to a clinician about whether or not lumbar puncture should be performed? A. Yes, that’s my view.”
Dr Knight, however, did not consider that the absence of papilledema was a reassuring sign, with respect to proceeding with a lumbar puncture, or, as he corrected himself, a reassuring absence of a sign. Dr Knight was not prepared to accept that the absence of papilledema or sixth nerve palsy, would inform the clinician that the presence of raised intracranial pressure was unlikely. He said that was not so in isolation and it depended on why the patient was undergoing a neurological examination. He said: “If the patient is asymptomatic and has no neurological signs then that might be so but if the patient has a history suggestive of raised intracranial pressure then I don’t believe those signs are sufficiently sensitive to say that the risk of intracranial hypertension is low.”
Both Dr Knight and Mr Wallace were of the opinion that, although papilledema was a well-recognised sign, it was an unreliable sign, in the sense that its absence did not mean there was no intracranial pressure. Likewise, the absence of sixth nerve palsy did not mean there was no intracranial pressure. These views were in contrast to Dr Allen’s opinion that the absence of those signs would have significantly reduced any concern of intracranial pressure.
Mr Johnston gave evidence as to what would have occurred, had the tumour been diagnosed at 30 December 1990, or 11 January 1991, as follows: “Q. If you assume that there had been a diagnosis made first on 11 January? A. Well I definitely don't think that would have made any difference.

Counsel for the respondent related the following aspects of the respondent’s history from 11 January 1991: that there was no relevant history of fever; that Dr Blissta found no meningeal signs; that on 12 January, the respondent had had an unsettled night and had vomited three times and was given Panadol with minimal effect; that on 12 January, she was still vomiting and afebrile; that the appellant had made a note, “some neck guarding probably no stiffness”; that the respondent continued to complain of headache in spite of having been given Panadol and that she vomited twice; that on 13 January she awoke feeling nauseated, had a small vomit and was given medicine for complaint of headache. Counsel suggested that given that clinical position, the diagnosis of meningitis was no longer appropriate on 13 January. The appellant disagreed.
The trial judge found that the appellant’s provisional diagnosis was an understandable one at the time it was formed, subject only to the history of headaches. It was that history that his Honour found, at [132], to be the troublesome feature in the approach taken by the appellant to the formation and maintenance of his provisional diagnosis.
It was relevant to his Honour’s consideration whether the appellant had breached his duty to the respondent that the appellant was aware, from his first consultation, that headaches were a feature of the respondent’s presentation. (11) The asserted undue reliance on non-paediatric specialists
The appellant contended that his Honour erred in finding that the appellant had breached his duty insofar as that finding was based upon the opinions of Dr Knight, Dr Hopkins, Mr Wallace and Mr Klug. It was submitted that those doctors expressed their opinions on the basis of their own specialisation and that the approach each would take would have been different from that of a general paediatrician.
Dr Knight said that at the time he first came to review this matter, he was significantly involved in general paediatrics and had attempted not to permit his opinion to be influenced by the ease with which he could investigate and intervene with patients. In this regard, Dr Knight agreed that one of his skills as a paediatric intensive care physician was the ability to react urgently in situations that were perceived to be putting a patient in immediate risk of their lives. He agreed that as his expertise was a paediatrician working in intensive care, his viewpoint was conditioned to treating patients who were “potentially in extremis”. He said it was customary for him to react in a way that ensured that tests such as CT scans were done as early as possible, when there were sufficient signs to indicate their use. He also agreed that a clinician such as the appellant admitting a patient a) for investigation and b) to treat dehydration, might adopt a slightly different course, exemplifying “the conservative or wait-and-see approach”. He agreed his report was read with those two factors had to be kept in mind, that is, the focus of his experience in emergency paediatrics and his recognition that clinicians in the position of the appellant in practice behave in a more conservative way.
Mr Wallace, likewise, accepted, at least inferentially, that his opinion was firmly based within his speciality. He accepted that if there was a solid diagnosis, in a clinical setting, of meningitis and encephalitis, it would be safe to do a lumbar puncture. He was then asked what the position would be if the diagnosis was only provisional. Mr Wallace said: “I was really thinking of this specific case. The thing that sticks in my mind about this case is that there’s headache and vomiting and that’s come within two days of the exposure to chickenpox and, therefore, we’re still in the incubation period of chickenpox and it’s so easy, with hindsight every child I see with headache and vomiting has got a brain tumour, so I don’t see the hundreds of patients that go in with other things such as migraine and headache and vomiting …”
Mr Klug also said that in reviewing the material and expressing his opinions in the case, his focus was on his area of speciality, that is, neurosurgery, and thus on the types of conditions that came to his attention. He said: “I certainly approached this problem as a neurosurgeon based on my experience. I would agree that a paediatrician with a wider knowledge of childhood diseases would perhaps consider it differently to what I did or what I would do.”
Dr Hopkins did not give evidence to the effect that his opinion was predicated upon the specialist approach he would take a neurologist. Nonetheless, the appellant relies upon the fact that he is a neurologist and not a paediatrician.
These concessions by the witnesses were no doubt considered by his Honour. (12) Conclusions on breach of duty: 13 January 1991
The question for his Honour’s determination was whether the appellant failed to exercise reasonable care in his treatment of and advice concerning the respondent. The standard of care was that of the ordinary skilled person exercising and having that special care: Rogers v Whitaker and Rosenberg v Percival.
Although the evidence before his Honour was lengthy and complex, in the end, the question became whether the appellant was negligent when, on 13 January, he did not respond to the neurological incident that the respondent had at 11 am on that day, by ordering a CT scan. For the purposes of resolving this question, we put to one side the issue on the cross-claim as to whether the appellant was negligent at an earlier point of time and assume that there was no negligence in the appellant’s initial provisional diagnosis and approach to the respondent’s treatment. We also put to one side the rejected case of the causative role of the lumbar puncture on 13 January.
As the above review of the evidence reveals, the respondent came under the appellant’s care in circumstances where she was very ill and he had been told by the previously treating paediatrician, Dr Mansour, that she had had chickenpox. Dr Mansour did not give him the history of the previous hospitalisation or any other history. Relevantly, he did not give him any history relating to headaches, nor particularly, headaches from 18 December. The respondent’s condition was acute and she needed immediate treatment for her severe dehydration. Given the history of chickenpox, the appellant considered that her symptoms were consistent with meningitis and possibly encephalitis. The respondent considered it necessary to confirm (or deny) that diagnosis by a lumbar puncture that he directed should be undertaken. A lumbar puncture is the correct diagnostic test for meningitis/encephalitis.
It is apparent that the appellant was concerned with the respondent’s neurological state, because he undertook a neurological examination when he first saw her in his rooms on 11 January, specifically, to see whether there was evidence of raised intracranial pressure. There was none, and the appellant’s focus was then on ascertaining the reason for the respondent’s deterioration in the preceding days. His concern was, and remained, with her headaches vomiting and general condition.
The appellant considered that, notwithstanding the respondent’s condition from the time of her admission to hospital up until the neurological episode on 13 January, her presentation was consistent with post-chickenpox encephalitis. There is no criticism of the respondent’s provisional diagnosis by any of the medical experts.
The appellant considered that the neurological episode on 13 January classically indicated that something was “going on” in the respondent’s brain. He also accepted that the signs of unequal pupils and the symptoms of increased headache and drowsiness were diagnostic of raised intracranial pressure. However, he did not consider that it was necessarily indicative of raised intracranial pressure. He also considered that raised intracranial pressure could be caused by meningeal encephalitis.
It is apparent from the appellant’s evidence that having made a provisional diagnosis of meningeal encephalitis, he saw nothing in the respondent’s condition up to and including the neurological incident on 13 January and prior to the seizure on 14 January, to reconsider that diagnosis, or to look for any other cause for her condition. The question to be answered is whether, in doing so, he failed to exercise reasonable care in the treatment of the respondent.
On the assumption on which we are proceeding, the appellant’s provisional diagnosis was reasonable. Accordingly, the question needs to be refined, to ask whether there was a lack of reasonable care in the appellant failing to reconsider his provisional diagnosis and take steps to ascertain whether there was any other cause for the respondent’s headaches, or any cause to explain the neurological episode on 13 January. Those two matters cannot, however, be disconnected, for reasons that we explain below.
The evidence in relation to the headaches was eventually all one way. In saying that, we put aside the evidence of Dr Mansour, which was not accepted by the trial judge on this question and not relied upon by either party on the appeal. The appellant did not assert that the headaches dating from 18 December were related to a complication of chickenpox. Nor did Dr Allen. The appellant’s position was that the headaches with which he was concerned on 11 January, continuing through to 13 January, were consistent with such a complication. Dr Allen said that it he would have viewed the history as being consistent with successive or composite viral illnesses, the second being chickenpox related, the first being overtaken by the second.
The evidence of Dr Knight, Mr Wallace and Dr Hopkins was that the history of the headaches in this matter was critical. Mr Wallace considered that the appellant had “missed the point”, in failing to take critical heed of the history of headaches, especially after the neurological episode on 13 January. Notwithstanding that no papilledema was detected at that time, Mr Wallace considered that chickenpox did not explain the history of headaches and their cause needed to be determined before anything else. Dr Knight and Dr Hopkins were of the same opinion.
The appellant’s clear understanding of the importance of the existence of a neurological condition is important in assessing what a reasonable practitioner exercising the ordinary skill of a doctor with his specialisation would do in the circumstances of this case. Mr Wallace stated that the history of headaches and vomiting from 18 December 1990 were indicators of possible raised intracranial pressure. The neurological episode on 13 January was a more focal indication of that possibility, notwithstanding the absence of papilledema. The medical evidence was that the absence of papilledema did not eliminate the possibility of raised intracranial pressure.
Even Mr Johnston, who otherwise supported the appellant’s case, said that if there were sufficient other symptoms, the absence of papilledema would not deter a clinician from performing a CT scan. The symptoms he identified as being sufficient included the history of 10 days of headache and vomiting prior to the chickenpox prodromal period. In this regard, Mr Johnston stressed that it was important, in making a decision to do a CT scan rather than a lumbar puncture, to confirm that there was an absence of any other explanation for the symptoms.
In this case, the appellant did not turn his mind to a consideration of any other alternative cause or explanation, either for the headaches from 18 December, or for the neurological incident on 13 January, or for the combination of both. His Honour did not make a finding that the appellant knew the history of the headaches. The evidence reviewed above explains why that finding was not clearly open. However, his Honour did find that the appellant should have known of the history of headaches. That finding was clearly open. If he did not know, as at 13 January, of the full history of the headaches, then he should have known, by taking steps more precisely to investigate or review the history. That history was readily available in the hospital notes, from the respondent’s parents and, for that matter, from Dr Sheiban, should the appellant have chosen to make such an inquiry.
The difference in the approach of the medical experts accepted by the trial judge and the approach considered to be appropriate by the appellant, demonstrates a quite fundamentally different diagnostic attitude. It was the appellant’s view that, having made an initial provisional diagnosis, that diagnosis should be pursued and other, differential, diagnoses not be simultaneously considered. He said that this was important so that the medical and hospital staff who had the continuing care of the respondent were not confused. His position was that it was then a question of considering the respondent’s progress having regard to the initial provisional diagnosis.
The alternative view was that obtaining a correct history was of fundamental importance. In this case, given the history of headaches dating from 18 December, some other provisional diagnosis should have been under consideration, although that did not have to be at the exclusion of the provisional diagnosis made by the appellant. The alternative diagnosis would have been that the headaches from 18 December were indicative of raised intracranial pressure. That being so, a more urgent response should have been made to the neurological episode on 13 January, which was also consistent with raised intracranial pressure. In short, whilst the neurological episode may have been consistent with a diagnosis of post-chickenpox meningitis, the history of headaches was not, and the neurological episode was not only consistent with post-chickenpox meningitis.
The evidence persuaded the trial judge that the appellant should have responded more acutely to the neurological episode on 13 January and ordered a CT scan at that time. He either knew, or should have known, of the respondent’s history of headaches, so that when the incident happened, he needed to assess not only the possibility of post-chickenpox meningitis, but also the fact there were two factors indicating possible raised intracranial pressure, namely, the pre-chickenpox headaches and the neurological incident. Whilst a CT scan and lumbar puncture were both diagnostic and not treatment procedures, the evidence was that in the case of raised intracranial pressure, time was of the essence. There was no evidence to the same effect in relation to the meningitis, which might be inferred in any event, given that the appellant, on the respondent’s admission on 11 January, was prepared to defer the lumbar puncture for a couple of days.
In our opinion, there was no error in the trial judge’s conclusion that the appellant breached his duty of care to the respondent on 13 January, nor in his reasoning process in arriving at that conclusion. The transcript reveals careful attention by the trial judge to the evidence, as is clear from his careful review of the facts in the judgment. The conclusion of negligence as opposed to error of judgment in a case such as this is a fine one to make. Examining the record, as we have sought to do, reveals the strengths of the competing views, but without the advantage of being present at the unfolding of the discourse amongst the experts and absorbing the manner and context of the expression of their views. The trial judge’s conclusion as to the failure of the appellant to respond more acutely to the events of 13 January was one founded in the acceptance of a significant body of (contested) medical evidence and a persuasive difficulty for the appellant as to his knowledge of the full history. We do not see any undue weight to the views of those without paediatric experience.
After an examination of the record and giving due weight to the advantages of the trial judge’s position and to his Honour’s careful analysis, we are not able to conclude that his Honour erred in his conclusion that the appellant was negligent on 13 January. (13) Conclusions on breach of duty: 11 January 1991
The respondent, by way of cross-appeal, contended that his Honour erred in failing to find that the appellant breached his duty of care to her prior to 13 January 1991. The premise of this challenge was that the diagnosis of post-chickenpox encephalitis was unsound from the outset, because it was fundamentally inconsistent with the history of headaches and vomiting.
The respondent’s challenge to his Honour’s finding was founded upon the following two essential considerations: first, the clear evidence that chickenpox related headaches would only have been experienced in the prodromal period, that is, approximately 24 to 48 hours prior to the chickenpox lesions appearing; and, secondly, the history of headaches, together with the history of vomiting, commencing from about 18 December. Accordingly, it was submitted that when the respondent presented to the appellant on 11 January, his diagnosis was unsound.
We have already discussed in detail the evidence as it related to the history of headaches. It is that history which is critical to the respondent’s submission. However, it was only Mr Wallace who was emphatic that the appellant’s diagnosis was wrong (see [111]-[121] above). Dr Knight expressed the opinion that the appellant should at least have considered a differential diagnosis and that such differential diagnosis should have included the possibility of an intracranial lesion (see [105] above). Much of Mr Wallace’s and Dr Knight’s evidence was directed to the question whether a lumbar puncture should have been performed on 13 January. There were two reasons that was a particular issue at trial. The first was that by undertaking the lumbar puncture, the appellant had not pursued a CT scan as he should have following the neurological incident on 13 January. The other was that the lumbar puncture itself was dangerous and there was a case run at trial that the lumbar puncture itself had caused brain damage.
In our view, there was no error in his Honour finding there was no negligence prior to 13 January. His Honour carefully assessed the evidence and it could not be said that the weight of the evidence was that the appellant was negligent from the outset. Rather, there was a variety of views on that and his Honour was entitled to have regard to the acute condition the respondent presented with on 11 January, her urgent need for intravenous fluids and the history the appellant was given at that time by Dr Mansour of the respondent having chickenpox. Further, his Honour did not make a clear finding as to when the appellant should have become fully apprised of the respondent’s history. It will be remembered that the appellant arranged for her immediate hospitalisation on 11 January, under the care of Dr Kohn. It appears that the appellant himself did not attend the hospital until 12 January. It was not suggested that showed any want of reasonable care. Accordingly, on one view, the earliest that the appellant might have had access to the records of the history of headaches was in the hospital notes, had they been reviewed by the appellant on 12 January.
The respondent also relied upon the fact that although the appellant, in his admitting note to Dr Kohn, stated that the respondent had clinical signs of meningitis, this was not replicated in the examination undertaken by Dr Kohn and Dr Blissta on 11 January. Dr Kohn found no neurological abnormality and Dr Blissta stated in his examination notes that there were no meningeal signs. The respondent contended that the absence of neurological signs in the one examination and meningeal signs in the other ought to have alerted, if not brought home, to the appellant that his provisional diagnosis of post-chickenpox encephalitis was dubious.
Even if that was the case, the weight of the evidence was that a conservative approach was not inappropriate, at least until the neurological episode on 13 January.
In any event, the respondent has not established that a CT scan should have and would have been carried out earlier than 13 January. There were medical views expressed to that effect, but they were in the minority. E. Causation: factual questions
At trial, two questions were raised on causation: first, whether the lumbar puncture caused the respondent’s deterioration on 14 January 1991; and, secondly, what were the consequences of the failure to seek to arrange for a CT scan on 13 January 1991, apart from the consequence that the lumbar puncture would not have been performed. It is only the second of these questions that is in issue on the appeal. (1) Likely treatment if a CT on 13 January 1991
His Honour concluded, at [300], that if the appellant had decided that there was an urgent need for a CT scan following the incident at 11 am on 13 January, such a scan would have been carried out on that day. His Honour considered that hospital protocol required that a neurologist or neurosurgeon be called and it was that doctor’s decision to determine if and when a CT scan ought to be performed. His Honour accepted the evidence of both Dr Maixner and Mr Johnston that if the case had been referred to either of them, a CT scan would have been arranged that day. Importantly, the evidence of Dr Maixner and Mr Johnston thereafter parted ways as to what subsequent treatment would have been given.
Dr Williams, who was on staff at the hospital in January 1991, gave evidence that it would have taken five to six hours from a decision being made that a CT scan was required for the scan to be carried out. If Dr Williams was aware of the respondent’s abnormal signs and her history, he considered that it would have been indicated clinically that an urgent procedure be carried out. Dr Maixner said that a scan could have been carried out on 13 January, which was a Sunday, if a decision had been made that the scan was urgent.
His Honour accepted that CT scans were routinely arranged by the registrar on duty, or on call, and that it was likely that the registrar would have been approached to arrange for the CT scan, rather than for a senior neurosurgeon to have been called in at that stage. In this regard, his Honour, at [300], accepted Dr Maixner’s evidence that: “…the likely scenario, because of the way the unit was run, [was that she] would have seen the patient, organised the scan, seen the results of the scan and conveyed the results of the scan to [the neurosurgeon].”

The appellant accepts that a CT scan could and would have been carried out on 13 January had a decision to that effect been made. The medulloblastoma would then have been discovered on 13 January.
At this point, it is critical to examine the evidence as to the treatment that would have been given to the respondent had the medulloblastoma been diagnosed on 13 January.
Mr Johnston stated that had the CT scan been done on 13 January and disclosed the medulloblastoma, he probably would have started the respondent on steroids and that such treatment would have improved her neurological condition, that is, the tumour-related brain swelling (brain oedema). However, he said that steroids are not particularly effective where the intracranial pressure is due to hydrocephalus.
Mr Johnston stated that it could not be said that steroids would have stopped the acute episode that the respondent experienced on 14 January because of their significantly reduced effectiveness where the pressure was not due to actual brain oedema. However, he considered that the introduction of steroids on 13 January would probably have had some beneficial effect. He said that treatment with steroids may have prevented a deterioration in the respondent’s condition, but that could not be said with any certainty. It was problematic as to what effect they would have and that it was “entirely possible” that even with close monitoring, the respondent would have deteriorated in the same way as she in fact did. Mr Johnston reiterated that although steroids would have been his selected treatment, it was “speculative” as to whether they would have had any role in avoiding (or minimising) the acute decline on 14 January, even on the assumption that the respondent would have been monitored by medical staff during that period.
Mr Johnston also said that even had a CT scan been carried out on 13 January, he would have not initiated any urgent procedure for removal of the tumour, but rather, would have pursued the course of a second ventriculostomy. He said there were certain disadvantages in a drain, stating, “[a drain] serves a purpose in an acute situation but in something slightly less acute a ventriculostomy would have been preferable”. That would have been performed in the ordinary course on the Tuesday or the Wednesday, that is, 15 or 16 January.
The trial judge accepted Mr Johnston’s evidence as to what he would have done in relation to the respondent’s treatment had she been referred to him on 13 January, that is, by treating her with steroids. His Honour also accepted that if a decision had been taken for a CT scan to be carried out on 13 January, it could have been performed later the same day, even allowing for the need for a general anaesthetic to be administered. His Honour also accepted evidence from Mr Wallace that in his opinion, there was an urgent need, on 13 January, for some measure to be taken to relieve the raised intracranial pressure.
Dr Maixner was of the opinion that, given the history of the incident of unequal, non-reactive pupils, with the respondent then having recovered from that episode, but still drowsy, she would have accorded the respondent high priority in relation to treatment. She considered that the history dictated that the respondent needed urgent relief from her severe intracranial hypertension. Dr Maixner said that the treatment would be dependent upon the surgeon. She said the simplest way to relieve pressure was to insert a drain to provide immediate relief from pressure, but that the respondent also needed removal of the tumour, which would almost invariably control the fluid build-up in the brain. Dr Maixner said that some surgeons would have undertaken the removal of the tumour immediately, whereas others would have waited. In either case, the respondent needed urgent relief of intracranial pressure.
The trial judge accepted that had Dr Maixner become involved in the respondent’s care on 13 January 1991, she would have treated the case as being urgent, but the surgical action to be undertaken would have been dependent upon the view taken by the surgeon.
Dr Klug would also have first administered corticosteroids for a patient who presented with medulloblastoma and secondary to it, hydrocephalus. Dr Klug explained that steroids are very effective in improving the situation of children with this disorder. Invariably, within 24 hours, corticosteroids would lead to substantial improvement in the condition and enable a fuller assessment of the patient. Dr Klug would then have planned a definitive surgical removal of the tumour, which would have also involved the insertion of a ventricular drain. If the child was in a good condition he would see no reason to insert a drain as a preliminary to operative treatment. He said it was only when a child presented in extremis with severe hydrocephalus when referred to him that he would, as a preliminary, insert a drain.
Dr Klug was asked if, in 1991, there were others in his areas of speciality that favoured the use of ventricular drains over the use of corticosteroids. He replied that it was not the policy in Melbourne (where he was in practice as that time). He reiterated that if the child presented in extremis with severe hydrocephalous, it would be a reasonable step to insert a ventricular drain to stabilise the hydrocephalic state. However, in his experience, that was a very rare requirement and the majority of patients with symptoms of raised intracranial pressure would be administered high dose corticosteroids. At the time of the operation for removal of a tumour, it was his invariable practice to insert a ventricular drain as part of that operation. Very occasionally, if the patient was in extremis, he would insert a ventricular peritoneal shunt prior to the operative procedure, rather than a ventricular drain.
Dr Klug stated that in Canada, the insertion of a ventricular peritoneal shunt for two weeks was standard practice for all children presenting with a posterior fossa tumour, as it was thought to achieve a simpler operative procedure, with less risk of complication. The shunt was a definitive procedure to drain the CFS internally, rather than by the use of an external bottle, as would be the case with the procedure favoured by Dr Maixner, and allowed control of the pressure for a two-week period to allow substantial improvement in the overall child’s condition, thereby minimising the risks inherent in the operative procedure.
Counsel for the appellant suggested that the shunt procedure was a “practice which existed” in Sydney in 1991. Dr Klug said he was not sure, but that there would have been a “small place” for a shunt to be inserted as a preliminary to operative treatment. As we understand that comment, the insertion of a shunt would not be a usual practice, but might occur in a small number of cases. He again explained that if the patient were in extremis from hydrocephalus, a preliminary intraventricular drain would have been the appropriate step to take. However, the vast majority of patients Dr Klug operated on with medulloblastoma did not, in his opinion, require a preliminary intraventricular drain, largely because the response to corticosteroids was so effective.
Dr Klug’s evidence recorded in the previous paragraphs was given in answer to a series of questions that commenced with the assumption put to him by the cross-examiner, that a CT scan had been done on 11 January and the medulloblastoma and hydrocephalus detected at the time. When asked if his response would be different if the tumour was discovered on 13 January 1991, Dr Klug stated that he would have to assess carefully the neurological condition of the child and if the hydrocephalus was extreme and potentially causing serious deterioration, a combination treatment may be used, namely, a ventricular drain and corticosteroids. (2) Findings of trial judge on likely treatment
At [299], his Honour stated that the question he needed to determine was how soon a CT scan could and would have been arranged, and what measures thereafter would have been put in place, and when, for the relief of raised intracranial pressure. His Honour said that what was relevant was how Mr Johnston would have acted if consulted on 13 January 1991, in circumstances where he did not share Dr Maixner’s view as to the urgency of the situation.
His Honour held, at [301], that he was not able to find as a probability that the course of treatment proposed by Dr Maixner would have been followed, had the CT scan been performed on 13 January, because of the conflicting evidence of Dr Maixner and Mr Johnston. Although there was a real chance that a decision would have been made to insert a drain as a matter of urgency, there were a number of contingencies which rendered that chance less than probable.
The chance that a drain would have been the course of treatment depended not only on Dr Maixner becoming involved in the respondent’s treatment that day, but also if she had, whether her view would have prevailed with the neurosurgeon who undertook the respondent’s case. It was clearly apparent on the evidence that had Mr Johnston become involved on 13 January, he would have taken a different approach to Dr Maixner. Dr Maixner conceded in her evidence that even if she disagreed with a neurosurgeon’s decision, she would generally have had to accept it.
His Honour considered that there was a “real chance” a decision would have been made to insert a drain as a matter of urgency. However, having regard to Mr Johnston’s status at the hospital, his Honour, at [304], considered that Mr Johnston’s evidence “reflect[ed] the likely neurological approach to [the] case at [this particular] hospital”. In other words, following the CT scan, it was likely that the respondent would have been treated in the first instance with steroids and not with the insertion of a drain.
His Honour concluded, at [306], that he was not persuaded on the balance of probabilities that if the appellant had called for a CT scan at 11 am on 13 January (as his Honour found he should have done), the discovery of the tumour and any consequential treatment would have avoided such brain damage as occurred on 14 January 1991.
His Honour thus rejected the respondent’s claim that she was entitled to damages for negligence resulting in brain damage arising from the harm she suffered on 14 January 1991, in circumstances where, on her case, the ultimate damage she sustained was not divisible.
His Honour found, however, that had the scan been called for urgently at 11 am on 13 January, it would have been performed that day, the medulloblastoma would have been detected and treatment would have commenced to reduce pressure, either by the prescription of steroids or by the insertion of a drain. His Honour considered that the use of drain would have given the respondent a better outcome, but it was necessary to heed the superior role Mr Johnston would have had in the decision-making process.
His Honour found that the respondent was deprived of the chance of being afforded one or the other of these treatments as a result of the appellant’s breach of duty. The respondent was therefore deprived of the chance of a better outcome, for which she was entitled to compensation. In valuing that lost chance, his Honour took into account the time lost in carrying out the EEG and performing the CT scan on 14 January. This was important on his Honour’s reasoning, because, as Mr Johnston acknowledged, “the earlier the intervention by neurosurgeons the better the likely result”.
As part of the appellant’s argument that the trial judge should have found that no damage was caused by any failure of due care on his part on 13 January, the appellant contended that the trial judge should have found that the course of treatment on 13 January after the CT scan would have been the administration of steroids.
We have already referred to the trial judge’s finding, at [304], that Mr Johnston’s treatment regime was the likely neurological approach to the respondent’s treatment had the medulloblastoma been discovered on 13 January. His Honour confirmed, at [378], that it was “more likely that steroids would have been prescribed rather than the placement of a drain”. The appellant contends that even if this finding was not a finding on the balance of probabilities, such finding should have been made.
The appellant also contended that his Honour should have found, in accordance with Mr Johnston’s evidence, that the effect of that treatment in reducing or minimising the effects was entirely speculative. Being entirely speculative, it fell outside the range of circumstances in which a court was permitted to award compensation for the loss of a chance, if such approach was permissible in any event.
Because there is a question as to the status of the finding which was made, and an issue as to the finding that should have been made, the finding itself needs to be put in context (including by referring in full to the finding at [304]).
Although it was not certain that Mr Johnston would have been the neurosurgeon consulted on 13 January, his Honour stated, at [304]: “However, in the absence of any evidence from some other neurosurgeon who may have been called in to treat the plaintiff, and having regard to Mr Johnston’s status at the hospital, I treat his evidence as reflecting the likely neurosurgical approach to the plaintiff’s case at this particular hospital.” (Emphasis added)
This finding led to the commencement of his Honour’s consideration as to whether the respondent was entitled to damages on the basis of a loss of the chance of a better outcome. His Honour undertook this consideration by reviewing the respondent’s course over the remaining time of her treatment. He found, at [333], that the brain damage occurred on 14 January, probably associated with increased cranial pressure that may have been avoided, or reduced, had some reasonable attempt been made to relieve the pressure before the deterioration occurred. The appellant accepts this finding.
Notwithstanding his Honour’s findings at [304] and [378], it is apparent from his Honour’s remarks at [306] and [333], that he was approaching the matter on the basis that had either form of treatment been administered, there was a possibility of a better outcome. However, in our view, his Honour’s finding as to the likely treatment that would have been administered (steroids) was a finding that he either made, or should have made, on the balance of probabilities. His Honour was satisfied that had the tumour been diagnosed on the CT scan on 13 January, treatment would have been commenced. There were effectively only two choices of treatment, steroids or a drain. A finding that steroids was the more likely treatment can only be a finding that, more probably that not, it would have been the treatment administered. This conclusion will be relevant to the question whether, if a plaintiff is entitled to damages assessed on the basis of a loss of a chance, his Honour’s approach to the assessment of that chance was correct. The anterior question is, however, whether damages may be awarded for the loss of a chance of a better medical outcome.
The trial judge’s conclusion expressed at [306] of his reasons was that the respondent had failed to persuade him (and thus prove on the balance of probabilities) that had the appellant exercised due care the brain damage that occurred on 14 January would have been avoided.
The respondent challenged this causation finding. To put the finding in context and to understand what is necessarily the full content of the finding it is helpful to deal, at this point, with loss of a chance, and, in particular, the factual analysis undertaken by his Honour. We will revert in due course to the legal legitimacy of this latter approach. (3) Lost chance of a better outcome
His Honour, at [353] ff, dealt with whether there was an entitlement to damages for the loss of a chance.
His Honour, considered himself bound by the decision in Rufo. He considered the case was one where the chance lost was less than a 50 per cent chance. His Honour said that in order to recover damages for the loss of this chance, the respondent had to prove that there was a real chance she would have had a better outcome had the negligence in treatment not occurred. She also had to prove that if she had been offered the opportunity of undergoing a CT scan and treatment subsequent to that on 13 January, she would have accepted it. His Honour made findings to that effect.
His Honour then found, at [377], on the probabilities: “(i) that the failure to relieve the [respondent’s] intracranial pressure during the twenty-four hour period prior to the [respondent’s] decline on 14 January 1991 was causative of brain damage occurring at and following the time of the [respondent’s] observed decline on 14 January;

It was submitted that the only factor his Honour had available to him in analysing the loss of a chance of a better outcome was a chance that steroids, if prescribed on 13 January 1991, would have avoided the deterioration the next day. However, the evidence did not support a finding that that chance gave rise to a compensable loss. Rather, it was submitted that Mr Johnston’s evidence was that it was not a reasonable supposition that the steroids would have stopped the episode on 14 January and it was speculative as to whether there would have been any better outcome.
If that was the only evidence, the respondent 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 arisen above there being a speculative chance.
However, Dr Klug’s evidence provided some support that the chance was more than speculative. He said that in non-acute conditions, the administration of steroids was very effective. However, he said that as at 13 January, he would have had to make a careful neurological assessment. If hydrocephalus were extreme, he would have considered treatment both by way of steroids and a drain. 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.
Although this evidence provides some support for the respondent’s claim, it provides little by way of assistance in assessing the loss of the chance. At the most, on this evidence, it can be said that the respondent lost some chance of a better outcome which ranged between speculative and some effect. Stating the matter in that way demonstrates the difficulty with such an assessment process.
What this evidence does illuminate, however, is the legitimacy of the trial judge’s conclusions as to the lack of evidence on the balance of probabilities that the damage suffered on 14 January by the seizure and deterioration would have been avoided had the appellant acted with due care and ordered a CT scan on 13 January.
There is, also, in our view, an error in the approach of the trial judge to the assessment of loss of a chance (if that approach be legitimate). It is clear from his Honour’s analysis that in reaching 40 per cent 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. On this basis, the loss of the chance of a better outcome considered (as his Honour did) as the damage that in fact occurred in these circumstances is considerably less than 40 per cent. If we were forced to place a percentage on it for the purposes of valuation we would be loathe to assess it as greater than 15 per cent. Thus, if we are wrong in the views that we express later about Rufo and a loss of chance analysis is legitimate we would consider that the respondent 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. (4) Respondent’s attack on the causation finding
The respondent’s attack on the trial judge’s finding at [306] of his reasons took two principal forms. (a) possible insertion of a drain
The first involved the proposition that his Honour should have concluded that on 13 January after the CT scan results were available Dr Maixner’s views would have prevailed and a drain would have been inserted in the respondent’s head, thereby immediately relieving the intracranial pressure and so, it was argued, avoiding the seizure and deterioration of 14 January.
We reject this argument. Much of its force rested on the proposition that Mr Johnston would have followed Dr Maixner’s views if he was aware of her view of the results of the CT scan that hydrocephalus was present. The matter was not put to Mr Johnston in terms and thus the evidential foundation for the argument in that respect is weak. But more importantly, and destructively for the proposition, it is clear that Mr Johnston would have been aware of obstructive hydrocephalus from the nature of the tumour.
In our view, this challenge to the trial judge’s findings as to the likely (that is, probable) result of the interplay of views of Mr Johnston and Dr Maixner should be rejected. (b) shifting burden of proof
The second form of the causation challenge can be seen to be based on Betts v Whittingslowe [1945] HCA 31; 71 CLR 637 at 649 and passages from Chappel v Hart [1998] HCA 55; 195 CLR 232, at [93]. This argument was put orally on appeal aided by some short written submissions.
The argument was that a CT scan was necessary to examine the cause of raised intracranial pressure. Because the CT scan was not done on 13 January the reason for the intracranial pressure was not found and no treatment was given. The purpose of intervention (whether by steroids or drain) would have been to avoid the very kind of damage that occurred. It does not matter that the trial judge could not be persuaded as to the efficacy of steroids. The damage that occurred was the damage for which the duty was called into existence to prevent. It then fell to the appellant as defendant, as a matter of onus, to show that the brain damage would have occurred anyway, that is to rebut the benefit of this legal causation principle.
In aid of these propositions, the respondent referred to evidence of Dr Wallace as to improvement which could have occurred but for the delay and the general comments of Mr Johnston as to the beneficial effects of steroids.
We reject this submission. First, Dr Wallace’s evidence was given on an entirely different foundation than the relevant one here: that if a CT scan had been undertaken, the course of immediate treatment was the administration of steroids. The general expressions of view of Mr Johnston that the steroids would have been of some benefit do not address the question of whether the seizure and other decline would have occurred anyway. His view was that any benefit in that regard was speculative.
As to Chappel v Hart and Betts v Whittingslowe regard should also be had to Schellenberg v Tunnel Holdings Pty Limited [2000] HCA 18; 200 CLR 121 at [51]; Roads and Traffic Authority v Royal [2008] HCA 19; 82 ALJR 870 at [31] and [143]. In Flounders v Millar [2007] NSWCA 238; 49 MVR 53 at [4]-[38] Ipp JA (with whom Handley AJA agreed) examined Betts v Whittingslowe and the question of the asserted shifting evidential onus when the tortfeasor had created a risk of the kind of harm that in fact occurred. Ipp JA set out, at [8]-[28], the authorities in this Court and in other intermediate courts of appeal in Australia that material increase in risk was not to be equated with material contribution to the injury and that the plaintiff must establish that it was probable that the risk created by the tortfeasor came home. We do not repeat or set out Ipp JA’s analysis, with which we respectfully agree. The cases which establish the above propositions and discussed by Ipp JA were: Bendix Mintex Pty Limited v Barnes (1997) 42 NSWLR 307; Wallaby Grip v Macleay Area Health Service (1998) 17 NSWCCR 355 at [16]; E M Baldwin & Son Pty Limited v Plane (1998) 17 NSWCCR 434 at [77]; TC by his tutor Sabatino v New South Wales [2000] NSWCA 380 at [58] and [59]; Seltsam Pty Limited v McGuinnes [2000] NSWCA 29; 49 NSWLR 262 at [119]; Coca-Cola Amatil (NSW) Pty Limited v Pareezer [2006] NSWCA 45 at [8]; Van Den Heuvel v Tucker (2005) 85 SASR 512 at 530-31; Batiste v State of Queensland [2002] 2 Qd R 119 at 124. (Though, compare Amaca Pty Limited v Hannell [2007] WASCA 158 at [395] which appears to support a shifting evidential onus once there is proven a breach of duty followed by an injury within the area of foreseeable risk.)
The argument of the respondent that the evidential onus shifted once it was proven that a risk had been created by the breach of duty of a kind seen in the injury reflected the views of Lord Wilberforce in McGhee v National Coal Board [1973] 1 WLR 1 at 6-7 referred to by Kirby J in Chappel v Hart at 273 which is contrary to the above authority in New South Wales, South Australia and Queensland.
There is no mechanical legal rule of causation based on the creation of risk, whether founded on a shifting evidential onus or otherwise, as submitted by the respondent. The passage in Betts v Whittingslowe reflects the entitlement of the fact finder to gain assistance from the existence of the breach of duty, the kind of risk produced by that breach and the damage that occurred in drawing relevant inferences from these matters as to causation: cf Roads and Traffic Authority v Royal at [31] and [143]. (The above is not intended to detract from any operation of the principles in Watts v Rake [1960] HCA 58; 108 CLR 158 and Purkess v Crittenden [1965] HCA 34; 114 CLR 164; but as to which see Seltsam v Ghaleb [2005] NSWCA 208 at [101]-[112] and Luntz, Assessment of Damages for Personal Injury and Death (4th Ed) [1.9.14].)
It is also to be noted that these arguments of the respondent on the cross-appeal were to the effect that the trial judge should have found that the whole of the permanent damage caused by the seizure and deterioration on 14 January (that is the 25 per cent of the total disability) was caused by the negligence on 13 January. No case was run below or on appeal that the delay in diagnosis and the need to carry out a CT scan and an EEG on 14 January caused some divisible part of the damage attributable to the seizure and deterioration on 14 January.
The references by the trial judge to the evidence at [238(e)] might be thought to enliven that possibility. However, the evidence was general in its nature and there was no particularity as to what damage harm might have been caused by the delay. This is particularly apparent in the evidence of Dr Klug, who when asked “Would that delay have resulted in harm to the patient …”, said “I believe that it had the potential to further harm the patient”. He gave no more specific evidence on this issue other than to say (shortly before this) that the delay between the seizure occurring at 11.45 and the drain being inserted at 2.30 was ”of some significance”.
The context in which Mr Johnston gave the evidence to which his Honour refers also demonstrates that his concession that the earlier the intervention the better the outcome was also general. Mr Johnston did not accept that it was a reasonable proposition that the administration of steroids on and from 13 January would have stopped the seizure from occurring on 14 January. He said it was “entirely possible that … the same course of events would have occurred”, that is, the seizure followed by the emergency procedure to insert the drain. Mr Johnston did not consider that would have occurred much more quickly than in fact occurred on 14 January. He then agreed with the proposition that in the case of a neurological incident such as the respondent suffered on 14 January the earlier the intervention by neurosurgeons the better the likely result. It was suggested that even minutes could be critical and he said that he agreed with that as a general proposition. However, Mr Johnston was not asked whether all of the damage sustained in the deterioration on 14 January was caused by the negligent treatment on 13 January, undoubtedly for the reason that that case was not run.
It is to be noted that the Court was not taken to this material on the appeal. Whilst the trial judge to a degree took this material into account in the various factors brought to account in the loss of a chance analysis, there was no case run that part of the harm caused by the seizure could be viewed as divisible and caused by the delay in operative procedure on and after 14 January. F. Departing from earlier authority

To the extent that the trial judge was not satisfied on the balance of probabilities that the failure of the appellant to take reasonable care had caused particular physical harm, he was entitled and bound by existing authority to allow the respondent to recover damages for loss of an opportunity of a better outcome: see Rufo. Leave having been given to the appellant to reopen the correctness of that decision, it is necessary to consider whether this Court should depart from its own earlier authority. Before considering that question, it is convenient to identify the relevant principles.
Critical to those principles is a consideration of the position of an intermediate court of appeal within the structure of the Australian judicial system, together with the doctrine of precedent as it applies within that system. The Supreme Court of New South Wales originally derived its powers and procedures from the English system: see Charter of Justice 1823 (Imp), s 2 and, generally, A Twomey, The Constitution of New South Wales (2004) at 718 ff. Its present position must be understood not merely by reference to its English roots, but by reference to both State and Commonwealth constitutional structures. The Court exercises both State and federal jurisdiction; it applies the general law as modified by statute, both State and Commonwealth. It is in all respects subject to the authority of the High Court and, in their respective areas of legislative competence, the State and Commonwealth Parliaments. Nor can it act with indifference to the determinations of other intermediate courts of appeal operating in the integrated federal system. To the extent that the State and Territory jurisdictions operate separately with respect to their own statutory laws and practices, no such factor arises in the present case. (2) Power to depart: constitutional framework
The obligation of any court to follow a statement of relevant legal principle of a court above it in the judicial hierarchy is a matter of constitutional principle. How it achieved that status, historically, need not be considered for present purposes. Much will necessarily depend upon the existence of a modern statutory scheme for appeals and, at a practical level, the availability of reasoned decisions by appellate courts. On the other hand, the questions whether, and if so when, a court may depart from its own prior authority raise issues of a different order.
Discussion of principle in this respect will tend to reflect an understanding of the court’s role within the constitutional structure of the country, including doctrines derived from the common law: see Wik Peoples v Queensland [1996] HCA 40; 187 CLR 1 at 182 (per Gummow J). Thus, the suggestion that a court might be bound by its own previous decisions, or by those of a court of co-ordinate jurisdiction, was once seen as reflecting a view that courts made the law. If courts merely “declared” the law and were fallible, they could err and a proper application of the law would require that error be acknowledged and corrected: N Duxbury, The Nature and Authority of Precedent (2008) at 40-41. As explained by Jessel MR in Osborne v Rowlett (1880) 13 Ch D 774 at 785: “Now, I have often said, and I repeat it, that the only thing in a Judge’s decision binding as an authority upon a subsequent Judge is the principle upon which the case was decided: but it is not sufficient that the case should have been decided on a principle if that principle is not itself a right principle, or one not applicable to the case; and it is for a subsequent Judge to say whether or not it is a right principle, and, if not, he may himself lay down the true principle. In that case the prior decision ceases to be a binding authority or guide for any subsequent Judge, for the second Judge who lays down the true principle in effect reverses the decision.”
In more sophisticated form, the principle invoked is that a court does not exercise a legislative function: see W Marshall, “Is a Judicial Tribunal, Either of the Last Resort or Otherwise, Bound by the Principles Laid Down by Itself on Previous Occasions?” (1860) 2 Juridical Soc Papers 331 at 345. This approach reflects an element of the separation of powers doctrine which can arise in cases where s 15A of the Acts Interpretation Act 1901 (Cth) is sought to be invoked to save a partly invalid Commonwealth law. One limit on the exercise of that power is that the court cannot be required to perform a feat “which is in essence legislative not judicial”: The King v Burgess; Ex parte Henry [1936] HCA 52; 55 CLR 608 at 676 (per Evatt and McTiernan JJ); and Pidoto v Victoria [1943] HCA 37; 68 CLR 87 at 109 and 111 (per Latham CJ).
After some vacillation, the House of Lords held in The London Street Tramways Company Ltd v London County Council [1898] AC 375 at 379 (per Earl of Halsbury LC) that “a decision of this House once given upon a point of law is conclusive upon this House afterwards”. Perhaps ironically, Lord Campbell, who was a strong proponent of the doctrine that prior authority must be followed, had invoked a separation of powers justification in Beamish v Beamish (1861) 9 HLC 274 at 339. It was, however, a view subject to heavy criticism: see, for example, J Landau, “Precedents in the House of Lords” (1951) 63 Juridical Rev 222 at 226 ff. Nevertheless, that approach remained until abandoned by the Practice Statement (Judicial Precedent) [1966] 1 WLR 1234. (For exposition of the approach now adopted in the House of Lords, see, for example, Regina v G [2003] UKHL 50; [2004] 1 AC 1034 at [30]-[35] (per Lord Bingham of Cornhill) and Horton v Sadler [2006] UKHL 27; [2007] 1 AC 307 at [29]-[31] (per Lord Bingham) and [39]-[41] (per Lord Hoffmann).)
Unsurprisingly, after 1898, the English Court of Appeal also considered itself bound to follow its own earlier decisions as reflected in the words of Buckley LJ in Produce Brokers Co (Ltd) v Olympia Oil & Cake Co (Ltd) (1915) 21 Com Cas 320 at 322: “I am unable to adduce any reason to show that the decision which I am about to pronounce is right. On the contrary, if I were free to follow my own opinion, my own powers of reasoning such as they are, I should say that it is wrong. But I am bound by authority – which, of course, it is my duty to follow – and, following authority, I feel bound to pronounce the judgment which I am about to deliver.”
The passage set out above from Produce Brokers was quoted in the judgment of Lord Greene MR in Young v Bristol Aeroplane Co Ltd [1944] 1 KB 718 at 724, in a judgment which relaxed the severity of the earlier approach, but only in limited respects. Lord Greene MR identified the three circumstances in which a previous decision of the Court of Appeal would not be followed by a later Court of Appeal as: first, where the later Court is confronted by two conflicting prior decisions requiring a choice to be made; secondly, where the later court comes to the conclusion that the earlier decision, although not expressly overruled, cannot stand with a subsequent decision of the House of Lords; and, thirdly, where the later Court comes to the conclusion that the earlier decision was given per incuriam.
In Australia, following federation, similar debates arose. The separation of powers doctrine, as entrenched in the Australian Constitution, did not provide any necessary answer to the question of whether the High Court had power to depart from its own earlier rulings and, if so, in what circumstances.
When the High Court was established, the Privy Council, unlike the House of Lords, was willing to reconsider its earlier decisions: see Read v Bishop of Lincoln [1892] AC 644 at 655. That approach was followed in Australian Agricultural Co v Federated Engine-Drivers and Firemen’s Association of Australasia [1913] HCA 41; 17 CLR 261 at 278-279 (per Isaacs J): “Our sworn loyalty is to the law itself, and to the organic law of the Constitution first of all. If, then, we find the law to be plainly in conflict with what we or any of our predecessors erroneously thought it to be, we have, as I conceive, no right to choose between giving effect to the law, and maintaining an incorrect interpretation. It is not, in my opinion, better that the Court should be persistently wrong than that it should be ultimately right.”
To the extent that this view may be seen to preclude considerations other than the preferred view of the legal principle, it has been abandoned: see Queensland v Commonwealth [1977] HCA 60; 139 CLR 585 at 599 (per Gibbs J) applied in Re Tyler; Ex parte Foley [1994] HCA 25; 181 CLR 18 at 38 (per McHugh J); discussed in A MacAdam and J Pyke, Judicial Reasoning and the Doctrine of Precedent in Australia (1998) at [8.21]-[8.37]. Importantly, the distinction between the power and the circumstances of its exercise was noted by the Privy Council in Geelong Harbour Trust Commissioners v Gibbs Bright & Co (1973-1974) 129 CLR 576 at 582, in the opinion delivered by Lord Diplock: “The High Court of Australia has always possessed the power, which the House of Lords itself only assumed as recently as 1966, to refuse to follow its own previous decisions if it thinks fit. This power however has been used but sparingly. The decision whether or not to exercise it is, in their Lordships view, one of legal policy into which wider considerations enter than mere questions of substantive law. The fact that the court considers its previous decision to have been plainly wrong is a prerequisite to discarding it, but it is by no means a decisive reason for doing so.”
At least with respect to the High Court, there may be doubt as to whether the conclusion that an earlier decision is wrong, if by that is meant erroneous in terms of the law as it stands, is indeed a precondition to departing from earlier authority. There are circumstances in which the Court may think that the law should be changed, usually because of changed social circumstances or changes in related parts of the law: see, for example, Australian Consolidated Press Ltd v Uren (1967) 117 CLR 221 at 241 (PC); Public Transport Commission (NSW) v Perry [1977] HCA 32; 137 CLR 107; Burnie Port Authority v General Jones Pty Ltd [1994] HCA 13; 179 CLR 520; Brodie v Singleton Shire Council [2001] HCA 29; 206 CLR 512. However, at least where the High Court has addressed an issue, this Court has no power to depart from such authority for any policy reason short of statutory amendment to the law: see Jacob v Utah Construction and Engineering Pty Ltd [1966] HCA 67; 116 CLR 200 at 207 (per Barwick CJ), 217 (McTiernan, Taylor and Owen JJ agreeing).
The situation of an intermediate court of appeal, whether with respect to decisions of courts of co-ordinate jurisdiction within the national system, or in relation to its own earlier decisions, raises different issues. The issues are analogous to, though not identical with, those arising with respect to the highest court in a particular judicial system. Where the apical court considers itself bound by its earlier decisions, it is highly likely that intermediate courts of appeal will similarly treat themselves as so bound, so long as the earlier decision has not been overruled by a court above it in the hierarchy. Where the apical court permits itself a degree of freedom not to follow its own earlier decisions, it does not follow that intermediate courts of appeal will be accorded a similar freedom. (3) Criterion for engagement: nature and quality of error
The circumstances in which the High Court will depart from earlier authority includes the case where a former decision is “clearly wrong” and where “there are no circumstances countervailing the primary duty of giving effect to the law as the Court finds it”: Australian Agricultural Co at 278-279 (per Isaacs J). The rule adopted by the UK Court of Appeal in Young v Bristol Aeroplane was rejected by Dixon J in Attorney-General (NSW) v Perpetual Trustee Co (Ltd) [1952] HCA 2; 85 CLR 237 at 243-244 as “incompatible with the practice of the Court and … inappropriate”.
Nor was the narrow approach in Young v Bristol Aeroplane adopted by intermediate courts of appeal in Australia. Shortly after Young v Bristol Aeroplane was handed down, Jordan CJ noted in Bridges v Bridges and Hooper (1945) 45 SR (NSW) 164 at 172: “Being of the opinion that the decision of the Full Court in De Luca v De Luca cannot be reconciled with the principle or with earlier or later authority, I think that we should not be justified as treating it as good law. I recognise that a considered judgment of the Full Court should not be lightly disregarded; but there is no principle in force in New South Wales which constrains us, as the Court of Appeal has recently held itself to be constrained in England ( Young v Bristol Aeroplane Co ) to follow an earlier decision if we are satisfied that it is wrong; and I do not think that we should tie our hands by the introduction of such a principle.” (Citations omitted)
See also Bennett & Wood Ltd v Council of the City of Orange (1967) 67 SR (NSW) 426 at 430 (per Wallace P with whom Holmes JA agreed).
The circumstances in which an intermediate appellate court regards itself as free to depart from its own previous decisions have been regarded in the past as a subject of practice and procedure for the court in question to determine for itself: Nguyen v Nguyen [1990] HCA 9; 169 CLR 245 at 268-269 where Dawson, Toohey and McHugh JJ said: “The extent to which the Full Court of the Supreme Court of a State regards itself as free to depart from its own previous decisions must be a matter of practice for the court to determine for itself. …

It is no doubt important that the two paragraphs from Nguyen set out above be read together. The constitutional importance of the doctrine of precedent cannot be entirely at large within a national integrated legal system to the extent that each intermediate appellate court is entitled to determine for itself its own practice with respect to following earlier decisions. That is particularly so in circumstances where, as will be seen below, intermediate appellate courts are required to take into account, and in some circumstances follow, decisions of courts of co-ordinate jurisdiction. The first sentence in the latter paragraph cited above involves a statement of constraining principle; the second appears to identify the consequence which should flow from the application of the principle of constraint. Like similar statements in relation to prosecution appeals against sentence, the effects of applying the principle are likely to depend upon how it is understood by lawyers and courts.
Further, the proper approach for an intermediate court of appeal must be identified in a context which takes account of the different roles the court has to play. As noted above, the power of the court arises within a statutory and constitutional system of appellate courts. That appellate system is now entrenched in Chapter III of the Australian Constitution, particularly s 73. The result is that the High Court operates at the apex of the hierarchical system in Australia. As explained by Gaudron J in Commonwealth v Mewett [1997] HCA 29; 191 CLR 471 at 524 with respect to federal jurisdiction, echoing comments in Kable v Director of Public Prosecutions (NSW) [1996] HCA 24; 189 CLR 51: “There can be no doubt that the Commonwealth is a single nation. And when matters fall for determination in the exercise of the judicial power of the Commonwealth, they fall for determination by courts in an integrated Australian legal system, comprised of federal courts created by the Parliament and State courts invested with federal jurisdiction, with this Court at its apex. Given that the Commonwealth is a single nation and given, also, its integrated legal system, it is not strictly accurate to speak in terms of a court exercising federal jurisdiction in a State or Territory …. Rather, the jurisdiction of this Court extends throughout the Commonwealth, as does the jurisdiction of the Federal Court. … And federal jurisdiction is no different when exercised by State courts. They, too, exercise federal jurisdiction throughout Australia, not merely in its States.” (Citations omitted)
In the context of imposing liability on the Commonwealth, the role of Chapter III was also discussed by Gummow and Kirby JJ at 545-547. Further, once it is accepted, as expressed in s 80 of the Judiciary Act 1903 (Cth) since its amendment in 1988, that non-statutory law in the exercise of federal jurisdiction is “the common law in Australia”, the import of an integrated national legal system becomes significant more generally in relation to the hierarchy of appeals. The common law of Australia is, of course, not restricted to federal jurisdiction: see generally Lipohar v The Queen [1999] HCA 65; 200 CLR 485; John Pfeiffer Pty Limited v Rogerson [2000] HCA 36; 203 CLR 503 at [97]-[100], [161], [193].
While it is not in doubt that an intermediate court of appeal in Australia has power to depart from its own earlier authority, and not merely in the circumstances identified in Young v Bristol Aeroplane, discussion as to the basis of the power provides guidance for the circumstances in which it may properly be exercised. Thus, one purpose underlying the principle of constraint is to maintain certainty in the law. As further explained in Nguyen, at 270 in the joint judgment: “… rigid adherence to precedent is likely on occasions to perpetuate error without, as experience has shown, significantly increasing the corresponding advantage of certainty.”
Early in the life of the Federal Court, the approach to the question as to when a Full Court of that Court would depart from an earlier Full Court decision was laid down by the then Chief Judge (later Chief Justice) Sir Nigel Bowen and Forster J in a joint judgment in Chamberlain v The Queen [1983] FCA 78; 72 FLR 1 at 8-9 where their Honours said: “We do not regard this court as being bound by its previous decisions. However, we will normally follow an earlier decision unless convinced that it is wrong. It was argued that Duff’s case was wrongly decided. It was a closely reasoned decision. We are not persuaded that it is wrong. We consider we should follow it.”
In the jurisprudence of the Federal Court, on a number of occasions, the expressions “plainly” or “clearly” wrong have been used: see, for example, Transurban City Link Ltd v Allan [1999] FCA 1723; 95 FCR 553 at [26]-[31], especially [29]; New Zealand v Moloney [2006] FCAFC 143; 154 FCR 250 at [133]-[139]. This can be seen to have been influenced by the use of the expression “plainly wrong” by the High Court in Australian Securities Commission v Marlborough Gold Mines Ltd [1993] HCA 15; 177 CLR 485 at 492. However, it is clear from the discussion in Transurban that those adverbs “plainly” or “clearly” do not limit the circumstances of departure to those in which error is patent or obvious or easily perceived. Rather they bespeak the quality of the error or the level of conviction of error that must be perceived: cf Plaintiff S157/2002 v Commonwealth [2003] HCA 2; 211 CLR 476 at [13] (per Gleeson CJ). The reference in Telstra Corporation Ltd v Treloar [2000] FCA 1170; 102 FCR 595 at [28] to “patent” is to be understood in its context of the approach to statutory interpretation, over which minds might reasonably differ.
The approach in Chamberlain of the necessity for conviction as to the error of a previous decision of an appellate court can be seen enunciated by Gleeson CJ in Clutha Developments Pty Ltd v Barry (1989) 18 NSWLR 86 at 99-100, at 100G: “… it is generally accepted that before it is appropriate for an appellate court to overrule one of its own earlier decision[s] it must entertain a strong conviction as to the incorrectness of the earlier decision.”

Without intending to be exhaustive, the notion of “plainly” or “clearly” wrong has been used, in the sense discussed above, in a number of jurisdictions to deal with judgments of appellate courts of the same jurisdiction or co-ordinate jurisdictions: Tillman v Attorney-General (NSW) [2007] NSWCA 327; 70 NSWLR 448; TSL v Secretary to the Department of Justice [2006] VSCA 199; 14 VR 109; RJE v Secretary to the Department of Justice [2008] VSCA 265 at [48]; Pilcher v HB Brady & Co Pty Limited [2005] WASCA 159 at [24]-[25]; The Queen v White [1967] SASR 184 (reversed in 122 CLR 467, but not on this point); R v Hood [2005] QCA 159; 2 Qd R 54 at [43]-[45].
The above jurisprudence reflects the need for restraint in departing from previous decisions in order to foster stability and predictability in the law and the administration of justice. It is important in exercising this restraint to establish and maintain a clear standard for the reconsideration involved. This has been accentuated by the requirement for intermediate appellate courts to have regard to, and generally follow, decisions of other intermediate appellate courts in Australia. That principle, which was applied to the construction of uniform national legislation in Marlborough Gold Mines has been expanded to apply to the common law generally in Farah Constructions Pty Ltd v Say-Dee Pty Ltd [2007] HCA 22; 230 CLR 89 at [135], where the Court stated: “Intermediate appellate courts and trial judges in Australia should not depart from decisions in intermediate appellate courts in another jurisdiction on the interpretation of Commonwealth legislation or uniform national legislation unless they are convinced that the interpretation is plainly wrong. Since there is a common law of Australia rather than of each Australian jurisdiction, the same principle applies in relation to non-statutory law.” (Citations omitted)
In Harrison v Melhem [2008] NSWCA 67; (2008) Aust Torts Rep ¶81-951 at [188], after referring to the statement of principle by Gleeson CJ in Clutha, with respect to the basis of departing from an earlier decision of the Court, Mason P continued: “We recently concluded that similar principles should apply both when this Court is asked to depart from one of its own decisions that [is] in point and when this Court is confronted with a decision of an Australian intermediate appellate court of comparable jurisdiction with reference to substantially similar legislation ( Tillman v Attorney General for the State of New South Wales …). I was in dissent, but I regard myself as bound by that decision although I observe that the High Court has subsequently endorsed an arguably different test ( Walker Corporation Pty Ltd v Sydney Harbour Foreshore Authority [2008] HCA 5, 242 ALR 383 at [31]).”
The decision in Walker Corporation was concerned with the application of judicially enunciated general law principles relating to compulsory acquisition law, which was statutory in origin. At [31], the joint judgment of Gleeson CJ, Gummow, Hayne, Heydon and Crennan JJ referred with approval to a statement of McHugh J in Marshall v Director General, Department of Transport [2001] HCA 37; 205 CLR 603 at [62] in relation to the construction of a statutory expression which appeared in legislation in various jurisdictions, but to which differing interpretations had been given. McHugh J continued: “But that does not mean that the courts of Queensland, when construing the legislation of that State, should slavishly follow judicial decisions of the courts of another jurisdiction in respect of similar or even identical legislation. The duty of courts, when construing legislation, is to give effect to the purpose of the legislation. The primary guide to understanding that purpose is the natural and ordinary meaning of the words of the legislation. Judicial decisions on similar or identical legislation in other jurisdictions are guides to, but cannot control, the meaning of legislation in the court's jurisdiction. Judicial decisions are not substitutes for the text of legislation although, by reason of the doctrine of precedent and the hierarchical nature of our court system, particular courts may be bound to apply the decision of a particular court as to the meaning of legislation.”
The precise scope of this dictum remains to be elucidated. It is not, for example, directed to general law principles. Marshall was concerned with a line of English authority, with respect to a phrase found in the Queensland legislation, which the High Court declined to follow. Somewhat different principles may apply in circumstances where two Australian legislatures, largely contemporaneously, adopt similar or identical language in pursuit of a common statutory purpose. In such circumstances, the coherent development of the law within Australia would not be promoted by the courts of one jurisdiction adopting a different construction to those of another. On the other hand, Walker Corporation indicates that a more relaxed approach might be adopted in such circumstances than would be the case with respect to federal laws or national uniform legislation.
Whether the application of the relevant test has the same operation in relation to decisions with respect to statutory construction as with respect to matters of general law is perhaps a question of more than academic interest in some circumstances: see Duxbury at 23; and see in this context Telstra Corp v Treloar at [27]-[28]. Such considerations, however, need not detain us in the present case as the issue concerns one of general law principle. Nevertheless, it will be necessary to take into account the fact that there is support in the Victorian Full Court for the earlier decision of this Court from which departure is being considered.
In Transurban, at [31], Black CJ, Hill, Sundberg, Marshall and Kenny JJ said: “Beyond this principle, we do not think it possible, or even desirable, to formulate exhaustive criteria upon which this Court should act when asked to reconsider an earlier decision, for so much will depend upon the nature of the controversy, the strength of the arguments and the particular circumstances attendant upon the case.”
There is, however, a difference between formulating “exhaustive criteria” and seeking to identify a structure and relevant matters to be considered in reviewing earlier authority. The phrase “plainly wrong” (or any like phrase) tends to focus attention on the jurisprudential nature and character of the error and underlying principle or course of authority or the conviction as to the existence of the error.
As explained by Aickin J in Queensland v Commonwealth at 621: “The expression ‘manifestly wrong’ has many times been used to indicate a basis upon which a prior decision may be overruled. With great respect to those who have used it, the expression, used without some qualification or explanation, suggests a subjective criterion not easily applied to distinguish one opinion from another.”
The phrases “plainly wrong” or “clearly wrong” can be understood to focus on at least one or more of the following attributes of a ruling: (a) the fact of error is immediately (in the sense mentioned in [283] above) apparent from reading the relevant judgment;

In our view, the first possibility is liable to be highly subjective and should not be required, where the other two possibilities are satisfied. The existence of (b) and (c) is a precondition to the exercise of the power to depart from earlier authority. (4) Criteria for exercise: discretionary factors
The use of the semantic label, clearly or plainly wrong, consistently adopted in earlier discussion, should be seen as a necessary, but not sufficient, condition for departure from earlier authority. In deciding whether to depart from such authority, an intermediate court of appeal should, as the High Court does, have regard to a range of other considerations which impinge on the question.
Guidance may be obtained from the approach adopted by the High Court, in reviewing its own earlier decisions, whilst recognising the differences in circumstance applicable to that Court. In John v Federal Commissioner of Taxation [1989] HCA 5; 166 CLR 417 at 438-439, in the joint judgment of Mason CJ, Wilson, Dawson, Toohey and Gaudron JJ the following statement appeared: “There is no doubt that this Court has power to review and depart from its previous decisions. However, such a course is not lightly undertaken: see Queensland v The Commonwealth [(1977) 139 CLR 585, esp at 599, 602 and 620]. Although there is, in the words of Dixon J in Attorney-General (NSW) v Perpetual Trustee Co Ltd [(1952) 85 CLR 237 at 243-244], ‘no very definite rule as to the circumstances in which [the Court] will reconsider an earlier decision’, in The Commonwealth v Hospital Contribution Fund [(1982) 150 CLR 49 at 56-58], Gibbs CJ, with whom Stephen J [(at 59)] and Aickin J [(at 66)] agreed, specified four matters which in that case justified departure from earlier decisions. The first was that the earlier decisions did not rest upon a principle carefully worked out in a significant succession of cases. The second was a difference between the reasons of the justices constituting the majority in one of the earlier decisions. The third was that the earlier decisions had achieved no useful result but on the contrary had led to considerable inconvenience. The fourth was that the earlier decisions had not been independently acted on in a manner which militated against reconsideration, as had been the case in Queensland v The Commonwealth .”
Queensland v Commonwealth was concerned with reconsideration of a previous constitutional decision: the factors which were identified as persuasive in that context were set out by his Honour at 630, but are not of direct relevance here. Nevertheless, they provide guidance in a more general context. “The cases to which I have referred above show that some general considerations have emerged which assist in the determination of the question whether a previous constitutional decision regarded as erroneous may, or should, be overruled. The first is that there should be no inhibitions about overruling a decision, the error of which has been made manifest by later cases which however have not directly overruled it. The second is that the Court will be slow to overrule, or should refuse to overrule, cases which ‘go with a definite stream of authority’ and do not ‘conflict with well established principle’. The third is whether the prior decision can be confined as an authority to the precise question which it decided or whether its consequences would extend beyond that question. The fourth is whether the prior decision is isolated ‘as receiving no support from prior decisions and as forming no part of what in one metaphor is called a stream of authority and in another a catena of cases.’”
While reference to the approach of the High Court to its own earlier decisions provides useful guidance to an intermediate court of appeal, such considerations cannot be automatically transposed. For example, the High Court is the ultimate guardian of constitutional principle, subject only to the authority of the people of Australia, voting on a referendum under s 128 of the Constitution. Secondly, while, in non-constitutional matters, a ruling of the High Court may be varied by statute, all other courts in the judicial hierarchy are susceptible to correction by the High Court. Principles designed to limit the occasions on which intermediate courts of appeal may differ from each other with respect to the general law, provide no clear guidance as to the proper course for an intermediate court of appeal once such differences have arisen. A relevant consideration for an intermediate court of appeal is that one role of the High Court is the resolution of such differences, a criterion expressly recognised as one relevant to a grant of special leave to appeal: see Judiciary Act, s 35A(a)(ii).
Transparency of approach is advanced by the articulation of such factors, albeit with the qualification that no list will be exhaustive. The articulation of these criteria is relevant to the values of consistency and certainty and address the extent to which they may be adversely affected by overturning earlier authority. Underlying the value of consistency, is the principle that similar cases should be judged alike, equality before the law being an essential element of the rule of law: see Judge Tanaka in South West Africa Cases (Ethiopia v South Africa; Liberia v South Africa) (Second Phase) [1966] ICJR 4 at 305-306, cited by Gaudron J in Street v Queensland Bar Association [1989] HCA 53; 168 CLR 461 at 571. However, as explained by Cardozo J in The Nature of the Judicial Process (1921) 112-113, that principle is not all-embracing: “One of the most fundamental social interests is that law shall be uniform and impartial. … Therefore in the main there shall be adherence to precedent. … But symmetrical development may be bought at too high a price. Uniformity ceases to be a good when it becomes uniformity of oppression.”
The phrase “plainly wrong” (or any like phrase) tends to focus attention on the jurisprudential nature and character of the error and underlying principle or course of authority or the conviction as to the existence of the error. Factors considered by Aickin J in Queensland v Commonwealth and in the joint judgment in John, on the other hand, suggest that there are other considerations bearing on the question as to whether the earlier decision should be overruled, error having been exposed to the requisite degree of conviction. These considerations are properly invoked because they are relevant to the underlying principles of certainty, predictability and transparency upon which the theory of precedent is founded. As explained by Nettle JA in RJE (at [104]), there must be “compelling reasons” for departure from earlier authority, whether in the same court or in a court of co-ordinate jurisdiction, a phrase encompassing both jurisprudential and practical considerations. G. Loss of a chance: recovery in negligence

In Rufo, the appellant, at about age 13, was diagnosed to be suffering from a serious inflammatory condition (lupus), in which a characteristic rash is associated with widespread internal pathology, including kidney damage. The appellant came under the care of Dr Hosking, a paediatric immunologist, in early February 1992, having first been seen by another specialist paediatrician who had prescribed Prednisolone, a corticosteroid drug: [58]-[59]. On 24 August 1992, the appellant was diagnosed as suffering from vertebral microfractures, caused by the corticosteroid dosages she had been having up to that time: [62].
On 8 June 1992, Dr Hosking substituted a different drug, Dexamethasone, which remained her medication until after the fractures were diagnosed: [64]-[65]. The trial judge found, relevantly for the issue on the appeal in that case, that the failure to introduce a “steroid sparer” in conjunction with the changed drug regime was negligent. Notwithstanding the breach of duty, the trial judge found that the appellant had failed to prove that damage had resulted.
Strictly, the question whether damages for loss of a chance were available was not in issue in Rufo, as the parties had agreed both at trial and on appeal that such damages were recoverable in point of law. The question was whether a case for damages for loss of a chance had been made out: see [206]. On that basis, M W Campbell AJA, who wrote the principal judgment in the matter, held, at [405]-[406], that: “Adopting a robust and pragmatic approach to the primary facts of this case … it seems to me that, more probably than not, the excess of corticosteroid consumed after 10 June 1992 in the context of the osteoporotic and vulnerable state of the appellant’s spine caused the loss of a chance that the appellant would have suffered less spinal damage than she in fact did.”
Hodgson and Santow JJA, in separate reasons, agreed. In doing so, their Honours considered the basis for making such an award.
Hodgson JA observed, at [3], that: “… the evidence strongly supported a conclusion that the negligence materially increased a risk, which was otherwise very substantial, that fractures would occur; and that the occurrence of the fractures was a realisation of this total risk (as distinct from the increment to the risk created by the negligence). This gives rise to the question whether the appellant is entitled to be compensated for the loss of the chance that, but for the negligence, the fractures would not have occurred (or would not have occurred at the time or with the severity of their actual occurrence).”
Hodgson JA considered that if the appellant was to recover damages on the basis of a loss of chance, then the trial judge erred in holding that the increased risk was too speculative to justify an award of damages. His Honour said that “so long as such an increase is material, I think the Court is required to do its best to assess it”: at [3]. (We will return to the importance of equating loss of a chance with increase in risk.) His Honour considered that in such a case it was necessary for a plaintiff to prove, on the balance of probabilities, that some actual quantifiable loss occurred as a result of the negligence, in order to be able to recover damages. He noted that, at least in some circumstances, the loss of a chance of commercial value was regarded as an actual quantifiable loss: Sellars v Adelaide Petroleum NL [1994] HCA 4; 179 CLR 332.
His Honour then observed, at [5], that: “There is a real question whether the loss of a chance approach is applicable in medical negligence cases, where an adverse result is suffered, for which there was a substantial probability even without negligence, but negligence has materially increased that probability.”
After referring to Hotson v East Berkshire Area Health Authority [1987] AC 750 and Naxakis v Western General Hospital [1999] HCA 22; 197 CLR 269, both medical negligence cases, his Honour concluded that such an approach was available in medical negligence cases and stated the basis upon which that was so, as follows: “[9] It seems clear that, if avoidance of the loss in question would have depended upon the plaintiff taking a particular course of action, the plaintiff must prove on the balance of probabilities that, but for the negligence, the plaintiff would have taken that course of action. The plaintiff cannot be compensated for the loss of a chance that the plaintiff might have done so: Sellars (at 353). However, otherwise I think it is consistent with the principles established in Malec and Sellars to say that it is enough if the plaintiff proves, on the balance of probabilities, that he or she has been deprived of a valuable chance.

His Honour concluded, at [12]: “In my opinion, the evidence in this case strongly supports the view that medical science can do no more than assert that there was a very substantial risk of the adverse result, that this risk was materially increased by the negligence, and that the adverse result was the realisation of the totality of the risk; and provide some basis for quantifying the chance that the adverse result would have been avoided if the negligence had not occurred. Accordingly, in my opinion this is a case that can be approached as one involving a loss of a valuable chance.”
His Honour assessed that chance as being less than 50 per cent.
Santow JA observed, at [18], that the High Court had not as yet laid down definitively that the loss of less than a 50 per cent chance for a better medical outcome was compensable and that there were dicta in opposing directions on that subject: referring to Naxakis; Sellars; Malec v JC Hutton Pty Ltd [1990] HCA 20; 169 CLR 638; Laferrière v Lawson [1991] 1 SCR 541; Roberts v Ohio Permanente Medical Group, Inc 668 NE 2d 480 (Ohio SC, 1996); and Jorgenson v Vener 616 NW 2d 366 (SD SC, 2000).
His Honour also noted that the English decisions were against the proposition: see Gregg v Scott [2005] UKHL 2; [2005] 2 AC 176; and Hotson v East Berkshire Area Health Authority, that in Canada the trend had also been against that approach: see Laferrière v Lawson and that the approach in the United States had been variable. His Honour then explained both the policy reasons and the underlying legal principles that accommodated such a claim.
As to the policy reasons, his Honour said, at [23]: “The question, why should there be recovery for loss of a chance when less than even, turns on the nature of the duty of care imposed on the medical practitioner. It is a continuum, starting with diagnosis and the duty to advise the patient as to his or her treatment and then embracing the mode of treatment itself. Bringing to bear reasonable care and skill in advice as to risks is in order that patient consent is then informed by knowledge of the material risks of the treatment options advised. That duty operates in a context where the patient needs to make an informed decision as to the anticipated outcome. Then when it comes to managing the treatment, here that of a drug dosage regime, again reasonable care and skill is brought to bear directed to a successful outcome by maximising the chances of its achievement. This is necessarily within the scope of the patient’s own informed decision about agreed treatment as well as its practical availability. We may know for example in hip replacements that a titanium hip offers a better outcome than a steel one. But cost considerations may render it practically unavailable.”
His Honour accepted that Sellars may be seen to stand in the way of a claim for damages for loss of chance of a better medical outcome. However, his Honour considered that the reasoning in Sellars on causation and damages was reconcilable with allowing recovery of a less than even chance in medical negligence cases. His Honour stated, at [39]: “ Sellars was a Trade Practices Act (Cth), s 52, case involving misleading and deceptive conduct. In such a case, while there is a duty not to misrepresent or mislead, it could not be gradated in terms of chance. There is no corresponding duty as is imposed on a doctor with regard to bringing to bear reasonable care and skill in maximising the chances of a successful medical outcome.”
His Honour went on to state that in medical negligence cases causation had to be demonstrated on the balance of probabilities in two respects. First, the chance itself had to be proven to exist on the balance of probabilities. Secondly, the plaintiff had to establish, on the balance of probabilities that, if offered the chance lost, in terms of treatment, he or she would have elected to have that chance. His Honour considered that to be an important control mechanism. He said, at [40]: “It means that loss of a chance cannot be invoked where there is not a greater than 50 per cent chance that the patient, properly advised, would have undertaken the particular course of treatment or operation. That is so though we are here dealing with a hypothetical event.”
His Honour likewise considered that Malec detracted from the availability of damages for such a loss of chance. He said that in Malec it had been established on the balance of probabilities that some injury had been caused by brucellosis: at [43]. However, in Rufo, the other damage was the loss of a better outcome. His Honour noted there was no associated recoverable injury brought about on the balance of probabilities, adding: “But I do not consider that distinction matters. In Malec , nothing was made of that distinction. It should not matter here, either.”
His Honour, at [45], returned to the question of policy, and said that it would be: “… productive of injustice if the plaintiff were to receive 100 per cent of the loss where a chance or prospect exceeds 50 per cent (say 51 per cent) yet receive nothing at all if such loss were, say, 49 per cent.”
His Honour recognised that on traditional causation principles, a plaintiff who was able to establish a loss of a better outcome on the merest balance of probabilities was advantaged. His Honour considered that consistency required that in such cases, the result should not be all or nothing, but damages should be assessed reflective of the percentage of chance, whether it be above or below 50 per cent. Having said that, his Honour concluded: “[53] The present is a case where loss of a chance does afford a proper basis for compensation. It highlights the importance of considering the duty of care not in isolation, but in conjunction with a fair assessment of the difficult task facing a doctor having to choose between various options each with their own risks.

To consider whether Rufo proceeded on a correct basis, it is necessary to address it in its general law context.
The tort of negligence is conventionally described by reference to the existence of a duty owed to the plaintiff by the defendant, a breach of that duty on the part of the defendant, identifiable harm suffered by the plaintiff and a relevant casual connection between the breach and the harm. These various elements are not, however, to be treated as independent of each other and hence to be addressed entirely separately. The manner of defining a duty will frequently answer the question as to whether it has been breached in particular circumstances; the means of identifying the harm will frequently determine whether the relevant causal connection exists. Accordingly, the intellectual discipline of separating the various elements must be exercised with an eye to the overall direction. In the present case, if the harm suffered by the plaintiff is identified as her irreversible brain damage, the tortfeasor either substantially or materially contributed to it, or he did not. If, on the balance of probabilities, the Court is satisfied that he did materially contribute, he will be liable for the whole of the harm suffered. In conventional terms, where a plaintiff suffers personal injury as a result of the negligence of another, the injury must constitute physical or mental harm, in the latter case amounting to a psychological injury or psychiatric illness going beyond a disturbed mental state: Mount Isa Mines Limited v Pusey [1970] HCA 60; 125 CLR 383 at 394 (per Windeyer J); Civil Liability Act 2002 (NSW), s 33. Conventionally, where no injury had been suffered, no liability will arise.
The development of the law of negligence has witnessed an expansion not only of the scope of circumstances in which a duty may arise, a process described by Windeyer J in Hargrave v Goldman [1963] HCA 56; 110 CLR 40 at 63-66, but also in the sorts of harm which are held to be compensable. These have expanded in various ways (and subject to various control devices) to include not merely physical and mental injury, but also economic loss. An expansion in one direction inevitably gives rise to attempts to generalise the new “exception” to form the basis of a more broadly stated general principle. Thus, in Fairchild v Glenhaven Funeral Services Ltd [2002] UKHL 22; [2003] 1 AC 32, the House of Lords permitted a claimant who had been exposed to asbestos fibres in several places of employment to claim against all his former employers for the loss suffered as a result of contracting mesothelioma. Because the claimant could not have established on the balance of probabilities that any one employer materially contributed to the disease (which may have been caused by the reaction to a single fibre) the claimant should have failed against each. The apparent injustice of that result led the House of Lords to allow the claim to succeed against all, thus providing an express exception (avowedly based on a policy choice) to the rule that causation must be established with respect to each tortfeasor on the balance of probabilities. As noted by Lord Hoffmann in Barker v Corus UK Ltd [2006] UKHL 20; [2006] 2 AC 572 at [5], it was “only natural that, the dyke having been breached, the pressure of a sea of claimants should try to enlarge the gap”.
Fairchild did not address a consequential issue, namely whether each of the employers was responsible for the whole of the damage or only for that proportion which might properly be attributed to negligent exposure at its hands. In Barker, that question was expressly raised for determination. That in turn required an answer to the question whether the defendants were each liable for a single indivisible harm (mesothelioma) or whether each was liable for increasing the risk of contracting mesothelioma. If the latter, apportionment was appropriate; if the former, each was liable for the full loss suffered by the plaintiff, although no doubt, as joint tortfeasors, apportionment between defendants would have been appropriate. The majority view was explained by Lord Hoffmann at [35] in the following terms: “Consistency of approach would suggest that if the basis of liability is the wrongful creation of a risk or chance of causing the disease, the damage which the defendant should be regarded as having caused is the creation of such a risk or chance. If that is the right way to characterise the damage, then it does not matter that the disease as such would be indivisible damage. Chances are infinitely divisible and different people can be separately responsible to a greater or lesser degree for the chances of an event happening, in the way that a person who buys a whole book of tickets in a raffle has a separate and larger chance of winning the prize than a person who has bought a single ticket.”
Lord Rodger of Earlsferry adopted a different analysis, with a different result in relation to apportionment: “[68] Lord Hoffmann describes the issue as being whether the damage caused by a defendant in a Fairchild case is the creation of a risk that the claimant will contract the disease so that each defendant will be liable only for his aliquot contribution to that total risk which has materialised. But, at the hearing, Mr Stuart-Smith QC, on behalf of Corus UK Ltd (‘Corus’) was at very considerable pains to say that he was not advancing an argument along those lines. On the contrary, he accepted that in a Fairchild case the defendant was liable for causing the mesothelioma, not for causing the risk of developing mesothelioma: the damage was the mesothelioma, not the risk of developing mesothelioma. He eschewed any suggestion that the ‘gist’ of the tort was creating the risk of harm rather than causing harm to the victim. The approach now adopted by your Lordships is, accordingly, not one advocated by the appellants.

The importance of coherence with general principles of tort law, as identified by Lord Rodger in Barker, requires further consideration.
First, it is clear that the principle identified in Fairchild was designed to address a problem of insufficient scientific knowledge to identify the precise causative factor in a known past event (the contraction of mesothelioma). That is, on any view, different from the problem with respect to a hypothetical assessment of an event which, because of negligent failure to diagnose, did not take place. Secondly, it is clear that the principle in Fairchild operated only to the benefit of a plaintiff. Thus, no question arose of its application in circumstances where the plaintiff could demonstrate on the probabilities that the negligence of the defendant had caused his or her disease or injury. Thirdly, although it is commonly said that a tortfeasor takes a plaintiff as he or she finds the plaintiff (including any abnormalities or susceptibilities), that principle is not apposite in the case of a sick person seeking medical assistance in respect of an unknown illness where the duty of the tortfeasor is to take reasonable care to identify and treat the illness.
Once these considerations are put to one side, there is a real question as to why the loss of a chance approach should operate only to the benefit of a plaintiff who cannot establish on the probabilities that the defendant’s negligence gave rise to an adverse outcome.
If the analysis adopted in Rufo and in the present case in truth owed its origin to the UK caselaw with respect to mesothelioma, its application would be misconceived. Indeed, it would be doubly surprising in circumstances where this Court has not adopted the British approach in Fairchild: see, for example, Bendix, where Mason P said at 316: “The law does not equate the situation where the defendant had materially increased the risk of injury with one where he had materially contributed to the injury.”

None of what their Honours said was directed to proof that damage was caused by (including in the sense of material contribution) the breach of duty. Causation, including the place of any “but for” analysis therein, has always been held to be required to be proved on the balance of probabilities. In the resolution of many causation questions in the law of negligence, the process of fact finding, including the drawing of inferences, will involve the consideration of the assessment of a chain of circumstances with a factual postulate (the careless conduct of the defendant) removed. That does not mean, however, that this aspect of the resolution of causation is evaluated otherwise than on the balance of probabilities. Malec provides no support for the resolution of any hypothetical element in causation on a loss of a chance basis by an increased risk short of a conclusion on the balance of probabilities.
That the majority judgment in Malec was saying nothing about causation is apparent from the Court’s adoption, at 643, of the statement of Diplock J in Mallett v McMonagle [1970] AC 166 at 176: “The role of the court in making an assessment of damages which depends upon its view as to what will be and what would have been is to be contrasted with its ordinary function in civil actions of determining what was. In determining what did happen in the past a court decides on the balance of probabilities. Anything that is more probable than not it treats as certain. But in assessing damages which depend upon its view as to what will happen in the future or would have happened in the future if something had not happened in the past, the court must make an estimate as to what are the chances that a particular thing will or would have happened and reflect those chances, whether they are more or less than even, in the amount of damages which it awards.”
In recent times the uncertain foundation of claims for loss of a chance is reflected in an extrapolation of two cases involving loss of a commercial advantage or benefit: Commonwealth v Amann Aviation Pty Limited [1991] HCA 54; 174 CLR 64 and Sellars.
The first, Amann Aviation, involved a consideration of heads of damage flowing from a breach of contract and a loss of an opportunity to secure a renewal of the contract when it would have expired, absent unlawful termination. The second, Sellars, concerned the loss of a commercial advantage or opportunity resulting from misleading conduct in contravention of the Trade Practices Act 1974 (Cth), s 52(1). Such a loss was held to be loss or damage suffered by a person for the purposes of a claim under s 82(1) of that Act. In such cases, the need to prove reliance upon the misleading conduct involved the plaintiff proving that he or she would, but for the inducement of the misrepresentation, have taken up an alternative commercial opportunity. The next step is to place a value on the lost opportunity. As explained in Gates v The City Mutual Life Assurance Society Ltd [1986] HCA 3; 160 CLR 1 at 13: “The lost benefit is referable to opportunities foregone by reason of reliance on the misrepresentation. In this respect the measure of damages in tort begins to resemble the expectation element in the measure of damages in contract save that it is for the plaintiff to establish that he could and would have entered into the different contract.”
As Sellars has been seen (wrongly in our view) as permitting the development by analogy of recovery for loss of a chance in medical negligence claims, it is important to examine the reasoning of the Court with some care.
In Sellars, it was asserted that the misleading or deceptive conduct of an executive of Poseidon had caused Adelaide Petroleum to lose a commercial opportunity to negotiate with a third party in better circumstances than it ultimately did. This was said to be “loss or damage” for the purposes of the Trade Practices Act, s 82(1). The subject of analysis was the economic value of an opportunity to negotiate commercial benefits. Mason CJ, Dawson, Toohey and Gaudron JJ expressed, at 339, the relevant question before them as: “… whether, in the circumstances of this case, loss of an opportunity to obtain a commercial advantage or benefit amounts to ‘loss or damage’ within s 82(1).”
The analysis of the majority, at 349, recognised that where a contract contained a promise to provide a chance, and breach resulted in the loss of that chance, that loss was to be assessed in terms of value. Assessment in those circumstances will include vicissitudes and possibilities. The recognition of the loss of a commercial advantage or an opportunity as harm was identified in Fink v Fink [1946] HCA 54; 74 CLR 127 at 143 (despite the result of the case) and Amann Aviation at 349. Their Honours noted that the assessment of damages in tort as assessed by reference to possibilities could be seen in Malec: Sellars at 349-350. There was no need to confine this approach to the assessment of damages for personal injuries. Their Honours then turned to what they saw as the critical question: whether there was any relevant point of distinction between contract and tort, damage being the gist of the action in the latter, at 351.
Their Honours discussed Sykes v Midlank Bank Executor & Trustee Co Ltd [1971] 1 QB 113; Gates v The City Mutual Life Assurance Society Ltd and Norwest Refrigeration Services Pty Limited v Bain Dawes (WA) Pty Ltd [1984] HCA 59; 157 CLR 149, at 351-353, and stated, at 353: “It may be that Sykes , Gates and Norwest are to be treated as cases which turn primarily on the issue of causation which is ordinarily governed by the general civil standard of proof. The distinction between proof of causation and damages was emphasized in Hotson v East Berkshire Area Health Authority . There Lord Ackner stated that the first issue that fell to be determined was that of causation. This was to be determined on the balance of probabilities. Once liability was established, the assessment of the plaintiff's loss could proceed, taking into account any reductions arising from the uncertainty of future events. When the issue of causation turns on what the plaintiff would have done, there is no particular reason for departing from proof on the balance of probabilities notwithstanding that the question is hypothetical.”
In cases of professional negligence where the breach of duty causes a claim to be statute barred (being the next group of cases discussed by their Honours at 353-354: Johnson v Perez [1988] HCA 64; 166 CLR 351 and Kitchen v Royal Air Force Association [1958] 1 WLR 563) the loss is one of a potentially valuable right, a chose in action – a form of property.
Their Honours then referred to Takaro Properties Ltd v Rowling [1986] 1 NZLR 22 and Craig v East Coast Bays City Council [1986] 1 NZLR 99 and the loss of a commercial opportunity by the negligence of the tortfeasor: at 354-355.
The enunciation of principle by their Honours was limited and expressed in terms of the deprivation of a commercial opportunity at 355: “Notwithstanding the observations of this Court in Norwest , we consider that acceptance of the principle enunciated in Malec requires that damages for deprivation of a commercial opportunity, whether the deprivation occurred by reason of breach of contract, tort or contravention of s 52(1), should be ascertained by reference to the court's assessment of the prospects of success of that opportunity had it been pursued. The principle recognized in Malec was based on a consideration of the peculiar difficulties associated with the proof and evaluation of future possibilities and past hypothetical fact situations, as contrasted with proof of historical facts. Once that is accepted, there is no secure foundation for confining the principle to cases of any particular kind.

The last sentence quoted above may be thought to elide two different elements, namely proof that the pregnant mother would have availed herself of the opportunity of termination and proof that the opportunity had some value. The latter question, his Honour held, was to be determined by reference to “degrees of probabilities or possibilities” as approved in Sellars. His Honour noted that although Sellars arose in “a business context” the “same principle” would apply in the present context in determining “the value of the loss of opportunity as such”: 57C-D (original emphasis).
The significance of the alternative approach identified in CES v Superclinics was that a question arose as to whether termination of the pregnancy would have been lawful in the circumstances of the case and, if not, whether damages could be recovered and, if so, what damages. Even if the hypothetical opportunity had been provided, its lawfulness was not readily assessable.
Having acknowledged that a claim defining the damage as loss of opportunity was not raised on the pleadings, the discussion of that issue was entirely obiter. A significant issue (on which Kirby ACJ and the other member of the majority, Priestley JA, differed) was the kind of damages which could be recovered by the mother, and particularly whether those damages could include the cost of raising a child. Whether that issue could have been avoided by pleading the harm as a lost opportunity is doubtful. (The substantive questions raised in the case were subsequently determined in Cattanach v Melchior [2003] HCA 38; 215 CLR 1, without reference to the possibility of a claim based on loss of opportunity.)
The next discussion of such a basis of a claim in relation to personal injury appears to have been Qantas Airways Ltd v Cameron [1996] FCA 1483; 66 FCR 246. Again, the claim formulated by the plaintiff, as representative of a group, was not based upon loss of an opportunity. Nevertheless, Lindgren J noted, at 295B-C: “But even if a loss of opportunity had been particularised as the loss or damage in question, a group member would have had to prove, in order to establish causation of loss and thus liability, at least that probably he or she would have heeded the warning by attempting to take a course of avoidance and that the lost opportunity of avoidance had had some value: see Sellars ….”
Chappel v Hart was concerned with the negligent failure of a medical practitioner to warn a patient of the risk involved in a particular operation, where the risk materialised and the patient suffered physical injury. The question was whether the failure of the doctor to warn the plaintiff of the risk materially contributed to the injury suffered. The majority (Gaudron, Gummow and Kirby JJ) concluded that it did. In the course of his separate judgment, under the heading “Causation: general legal propositions”, Kirby J referred to this issue in the following passage at [93]: “9. Valuing a lost chance: A further way in which, in some circumstances, the difficulties of causation for a plaintiff are alleviated is by treating the plaintiff's loss as a ‘loss of a chance’. … In CES v Superclinics … I indicated my attraction to this approach as a more rational and just way of calculating damages caused by established medical negligence.”
His Honour said that such an approach had support of the High Court in Sellars, of the House of Lords in McGhee, and of the dissenting judgment of La Forest J in Laferrière v Lawson at 613. However, his Honour acknowledged that the weight of judicial writing in England and Canada was to the contrary. There is room for doubt as to the nature of the point raised by Kirby J in Chappel v Hart, as the reference to CES v Superclinics was directed to a means of calculating damages “caused by established medical negligence”, suggesting that causation and injury had already been demonstrated on the balance of probabilities.
The following year the High Court considered an appeal from a jury verdict for defendants in a medical negligence trial, the verdict being entered on direction of the trial judge: Naxakis. Gleeson CJ noted at [5], having determined that the appeal should be allowed and that there should be a new trial: “In that connection, there was some reference in argument to the possibility that the plaintiff’s case might be put on an alternative basis, relating to the loss of a chance. That question was not examined by the trial judge or by the Court of Appeal, and it was not fully argued in this Court. I express no opinion about it.”
Gaudron J considered the question in more detail at [28]-[36], stating at [29]-[30]: “It is well settled that, where breach of contract results in the loss of a promised chance, that is an actual loss for which damages will be awarded ‘by reference to the degree of probabilities, or possibilities, inherent in the plaintiff's succeeding had the plaintiff been given the chance which the contract promised.’ So, too, damages may be recovered for a commercial opportunity that is lost in consequence of a breach of contract. And it was held in Sellars … that the loss of a commercial opportunity is ‘loss or damage’ for the purposes of s 82(1) of the Trade Practices Act 1974 (Cth). Moreover, there is no reason in principle why loss of a chance or commercial opportunity should not constitute damage for the purposes of the law of tort where no other loss is involved. However, different considerations apply where, as here, the risk eventuates and physical injury ensues.

Her Honour noted that in a “failure to diagnose or treat” case the contention that the loss suffered is the loss of a chance rather than the injury or physical disability that eventuates “is essentially different from the approach that is traditionally adopted”: at [32]. Her Honour rejected that approach in such a case: at [36]. However, she did so in part upon the view that difficulties in establishing causation in such cases were exaggerated: at [31]. Whether her Honour’s views with respect to causation have been adopted in full may raise an issue in respect of that part of her Honour’s reasoning.
McHugh J made no reference to the issue; Kirby J referred to it but expressed no opinion in the circumstances of the case: at [77].
Callinan J took a different approach. His Honour stated at [128]: “On the evidence here the jury were entitled to hold that the failure, the treating doctor's omission either to undertake an angiogram or the failure to give any consideration to the undertaking of it, materially contributed to the appellant's condition. They would also be entitled to take an alternative view that the second respondent's conduct, although it might not be possible to say (on the balance of probabilities) that it definitely materially contributed to the plaintiff's final condition, at least caused him to lose a valuable chance (the value of which it was for them to assess) of avoiding being in the condition that he now finds himself. There is still, in my opinion, room for the operation of the loss of chance rule (particularly in cases involving the practice of what is even today said to be an art rather than a scientific skill), enabling a plaintiff to recover damages to be equated with, and reduced to the value of the chance he or she has lost, rather than the damages which would be appropriate if it has been proved on the balance of probabilities that the plaintiff's condition owes itself to the defendant's acts or omissions.” (Citations omitted)
His Honour acknowledged that there were difficulties with such an approach: at [129]. In particular, he expressed the view that it should only work in favour of a plaintiff: that is a plaintiff who had established loss on the balance of probabilities should not have his or her damages reduced, but the plaintiff who established a lower chance, so long as it was real and substantial, should receive a proportionate award: at [130].
As subsequent authorities have recognised, conflicting dicta in the High Court provide no guidance to intermediate courts of appeal, let alone trial judges, as to how to approach such an issue. Nevertheless, recognition that such a course would be a departure from the traditional approach and that it involves significant issues of policy should discourage both trial judges and intermediate courts of appeal from adopting such an approach, absent further guidance. More importantly, Callinan J did not reason by way of analogy from the commercial cases, such as Sellars, and Gaudron J, who expressly referred to Sellars, stated that “different considerations apply where … physical injury ensues”: at [29]. With respect, that must be so.
The earliest reported Australian case in which damages were expressly awarded for harm identified as a loss of a chance of a better medical outcome appears to have been Gavalas v Singh [2001] VSCA 23; 3 VR 404. That case also involved a failure to diagnose a brain tumour. The negligence, as found by the trial judge, resulted in a delay of some 10 weeks. Damages were awarded by the trial judge on the basis of a prolongation of intractable headaches and left-sided weakness during that period. There was, accordingly, no dispute that the tort was established. The question was whether the trial judge had failed to allow some further amount for the loss of a chance of a more successful removal of the tumour when the diagnosis should have been made. The primary opinion in the Court of Appeal was delivered by Smith AJA (Ormiston and Callaway JJA agreeing). Smith AJA noted debate as to whether damages for a lost opportunity were available in tort, noting “judicial support for the two extreme positions but no binding authority denying the entitlement to compensation for such a loss in all cases”: at [37]. His Honour continued, at [38]: “It is difficult to see any reason in principle why a plaintiff should not be compensated in appropriate circumstances for a lost opportunity if it flows from the alleged negligence. It has been said: ‘The law of negligence may be seen as directed to several purposes, but purposes of compensating the injured and promoting reasonable conduct are prominent among them.’

The first internal quotation set out by Smith AJA in the passage above was from the judgment of Hayne J in Chappel v Hart at [123]. However, that view did not lead to a conclusion of upholding liability on the facts of Chappel. The second passage quoted by Smith AJA, from Sellars at 355 (set out at [344] above], followed a discussion of the principle that hypothetical questions with respect to the assessment of loss should be assessed by reference to possibilities in hypothetical circumstances and applying Malec as set out. The next passage in the joint judgment, not quoted by Smith AJA, was as follows: “Furthermore, it is an approach which conforms to the long-standing practice of taking into account contingencies in the assessment of damages.”
In Gavalas it was not in question that the plaintiff had established liability on the part of the defendant, including the element of actual harm. On one view, the assessment of loss in such circumstances was properly approached on the basis of the principles established in Malec. Sellars did not suggest any departure from conventional principle in this respect. (4) Reconsidering ‘loss of a chance’ as a form of harm
In Rufo there was no need to consider the conceptual basis of such a claim because “there was no challenge to Studdert J’s formulation of what the appellant was required to prove to recover damages for loss of the chance of a better outcome”: at [207] (per M W Campbell AJA). Further, to the extent that his Honour relied upon English case-law in support of the conceded approach, that analysis cannot stand alongside the more recent analysis of that case-law provided in Barker.
As appears from the passages set out above, Hodgson JA and Santow JA approached the case on the basis that, as explained by Hodgson JA at [9]: “However, otherwise I think it is consistent with the principles established in Malec and Sellars to say that it is enough if the plaintiff proves, on the balance of probabilities, that he or she has been deprived of a valuable chance.”
To say that such an approach is consistent with Malec and Sellars is not to justify its adoption. Indeed, it is, with respect, to ignore the distinction between causation of loss and assessment of proven loss, established in Malec and the fact that “different considerations apply” in respect of medical negligence, as compared with the circumstances in Sellars, as noted by Gaudron J in Naxakis at [29]. Further, once one views the harm (and its assessment) by reference to the materialised physical harm or debility, the role of increased risk in the analysis becomes clear.
In State of New South Wales v Burton [2006] NSWCA 12 the plaintiff had suffered post-traumatic stress disorder as a result of being fired upon whilst acting as a police marksman at the scene of a siege. He did not sue the offender, but the State, alleging a failure on its part to provide appropriate counselling. Spigelman CJ stated at [25]: “The PTSD was not caused by the failure to have counselling, or take other ameliorative steps. It was caused by the shooting. The relevant loss in this case is the loss of a chance of a better outcome. The focus of attention must be on the extent to which the Respondent’s condition would or could have been improved had earlier counselling been provided.”
As explained by Lord Walker of Gestingthorpe in Barker, “indivisible damage is not always instantly recognisable”: at [111]. The law has long recognised that most forms of harm are capable of being the result of a combination of causal elements. However, the purpose behind a proposed division of the harm into constituent parts may affect the outcome of the exercise. Conventionally, with indivisible harm, the plaintiff can recover the full amount of his or her loss from any contributing tortfeasor. While there may be apportionment between tortfeasors, there is usually no apportionment between tortious and non-tortious causes, although, in assessing damages, the amount may be limited where it can be shown that a pre-existing cause would have achieved the same level of disability after a certain time in any event.
Different questions arise where the plaintiff cannot demonstrate that particular negligence would, on the probabilities, have caused any worse outcome for him. Putting to one side the failure to warn cases, such as Chappel v Hart, which does not itself support recovery on the basis of a loss of a chance, the result in a particular case may depend upon how the plaintiff seeks to identify the harm suffered and whether, as a matter of fact, some divisible element of harm attributable to the defendant can be established, as suggested in Burton.
To the extent that these authorities suggest that a chance of avoiding or diminishing the severity of a disease or injury is sufficient harm to satisfy the requirements of the tort of negligence, that is a step having potentially far-reaching consequences. For example, in what cases will it apply? Will it limit recovery in cases where the plaintiff establishes a loss to a degree of probability between 50 per cent and, say, 90 per cent? Perhaps because the issues have not previously been fully debated (such a claim was not pleaded in Gavalas, its availability was conceded in Rufo, which was followed without its validity being questioned in Burton) these issues have not been adequately addressed. As the UK courts have all recognised, they involve issues of high policy affecting the scope and operation of the most dominant aspect of tort law. A statement that “[n]o advanced system of law could deny recovery where late diagnosis, in breach of duty to the patient, appreciably reduces the prospects of success of an operation” (Gavalas at [15]) does not provide reasoned support for a major new direction in negligence. The policy involved in this departure from orthodoxy can be seen in the reasons of the Supreme Judicial Court of Massachusetts in Matsuyama v Birnbaum 890 NE 2d 819 (Mass 2008); as to which, see (2009) 122 Harvard Law Review 1247.
Similar difficulties have been identified in the literature with respect to French law, which, it has been said, recognises a lost opportunity of avoiding injury or realising a benefit as a form of compensable harm, at least where the risk eventuated: see L Khoury, “Causation and Risks in the Highest Courts of Canada, England, and France” (2008) 124 Law Quarterly Review 103 at 122. However, to define the harm as the loss of a chance of a better outcome will only allow an assessment of loss by taking a proportion of the harm in fact suffered, which casts doubt on whether the risk or loss of opportunity is in truth the harm for which compensation is provided: Khoury at 125-127.
Professor Stapleton considers that the loss of a chance cases might be more satisfactorily explained on the basis that the damage was constituted by the immediate economic loss to the market value of the asset that was the focus of the lost chance: J Stapleton, “Cause-in-fact and the Scope of Liability for Consequences” (2003) 119 Law Quarterly Review 388 at 406 ff.
Importantly for understanding how such an approach does not intrude into traditional causation principles, Professor Stapleton explained, at 410: “Using the ‘present-damage-to-economic-value’ formulation, we could neatly conceptualise cases where a solicitor’s negligence has resulted in a legal claim of C becoming statute-barred. Before the negligence C had a chose-in-action of a particular worth (which would be determined by the likelihood of success in the action). The tort destroyed this asset. The cause-in-fact stage can then be stripped of unorthodox loss-of-a-chance devices. The cause-in-fact task for C becomes: can C prove, on the balance of probabilities, that the tort destroyed the asset?”
The appellant submitted that Sellars was properly described as involving the determination of a “present damage to economic value” in the sense used by Professor Stapleton. Whether or not that is the best way to express the problem perhaps does not matter. What can be said, however, is that such an analysis may have value with respect to the loss of a commercial opportunity, but cases such as Sellars provide no basis to undermine the existing orthodoxy in the law of recovery of damages for personal injury as a result of negligence. To the contrary, the reiteration in Sellars of the statements of general principle as to proof of causation were contradictory to any permission to extend loss of a chance outside the realm of loss of a commercial opportunity.
In case of physical harm or injury, the damage can generally be seen to be the physical injury, debility, pain, bodily or mental harm or deterioration and consequences thereof. The court awards damages for that harm and its consequences. This occurs when the breach of duty is proven on the balance of probabilities to have caused the damage. Despite the different expression of the loss of an opportunity or chance, the damages in such a case are still formulated by reference to the harm that has materialised. This can be seen in Rufo itself and the approach of the trial judge in the present case. The damages awarded in the present case were for the physical harm suffered, but calculated, not on the basis of full recovery consequent upon causation having been proved on the balance of probabilities, but rather on a lesser scale commensurate with the increased risk caused by the breach of duty. This was consistent with the part played in the reasons of Hodgson JA in Rufo at [3] of the increase in risk (see [306] above). The “opportunity” that was lost was to have a better outcome without negligence, the value of the lost opportunity being assessed as the percentage risk or chance of the harm that did occur. So, here, 25 per cent of the whole debility was caused by the seizure and deterioration on 14 January. The respondent did not prove on the balance of probabilities that that damage was caused by the breach of duty on 13 January. Rather, it is said that because of the breach of duty on 13 January the respondent lost the opportunity of a better medical outcome (measured by the avoidance of the damage on 14 January) assessed at 40 per cent. This is to say no more than that the assessed likelihood of the damage is 40 per cent, so that the respondent has “lost the opportunity” of a better outcome (the harm not occurring) calculated accordingly. Yet, one thing is plain: the respondent was awarded damages not for the harm of suffering the increased risk, but by reference to the physical harm and debility suffered, which, ex hypothesi, had not been shown to have been caused by the breach of duty.
It is to be recalled that causation, in this context, incorporates the notion of material contribution: Bonnington Castings Ltd v Wardlaw [1956] AC 613; and Chappel v Hart at 244 (per McHugh J).
Once one appreciates that the so-called loss of an opportunity in this field is in fact the increased risk of harm (short of proof on the balance of probabilities) there is an essential conflict between the concept of damages for loss of a chance and the authorities discussed by Ipp JA in Flounders v Millar, in particular Bendix; and see now Royal at [25] (per Gummow, Hayne and Heydon JJ) and [135]-[144] (per Kiefel J). Though the debate in such cases as Bendix; Wallaby Grip; E M Baldwin v Plane; and Seltsam v McGuinness has generally been discussed (as M W Campbell AJA did in Rufo at [413]-[446]) in the context of causation and material contribution, once one appreciates the true character of loss of a chance in medical negligence cases as damages for increased risk of harm, one needs to confront and depart from cases such as Bendix. No such attempt was made in Gavalas or Rufo. Nor is there any warrant for this course to be taken. (5) Gavalas and Rufo not followed
Somewhat paradoxically, factors favouring caution generally may favour a departure from existing authority in particular circumstances, one of which arises in this case. Recovery for loss of a chance was accepted in Rufo for the first time in this Court without clear support from High Court authority; although its adoption was by no means inadvertent, it was not opposed by any party to those proceedings. Rufo was followed, but not challenged, in Burton. In neither case were the full ramifications of such a development analysed. Accordingly, it is appropriate to consider whether it was a development which this Court would now take if Rufo and Burton had not preceded this case and, if it would not, whether it should now depart from that authority. At the second stage of that analysis, it is appropriate to ask not merely whether the step taken in Rufo was plainly wrong, but also whether the step having been taken, the authority should properly be followed. The fact of consent may diminish the justification for adhering to a precedent held to be erroneous in principle.
There was no licence to be found for the course taken in Rufo in the dicta of single justices of the High Court: Callinan J in Naxakis at [127]-[130] and Kirby J in Chappel v Hart at 272, contra Gaudron J in Naxakis at [20]-[36], and Gleeson CJ in Naxakis at [5] expressing no opinion. Sellars stands as authority for the proposition that deprivation of a commercial opportunity or advantage is compensable harm in contract, tort and under the Trade Practices Act, its value to be assessed on the basis of possibilities as an hypothetical circumstance; but the general standard of proof will govern the issue of causation of loss, properly identified.
The arguments in favour of permitting damages for loss of a chance cannot, in circumstances where the approach conflicts with conventional authority, be justified on the basis that there is no express superior court authority inconsistent with the proposed approach. Nor can the approach be preferred on the basis that it avoids the potential injustice which may flow from a bright line drawn in accordance with the balance of probabilities. Such a rationalisation would have operation well beyond the scope of the proposed new principle.
It is for the High Court, and only the High 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. 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 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 at 232 (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 (by redefining what is “harm”) in personal injury cases.
Our view is supported by final courts of appeal; in the UK see Wilsher v Essex Area Health Authority [1988] AC 1074; Fairchild; Hotson v East Berskshire Area Health Authority; Gregg v Scott and Barker; and in Canada, see Laferrière v Lawson.
A further matter to be considered is that there is now a statutory definition in the Civil Liability Act which implicitly does not include a risk of physical or mental injury. Part 1A of the Civil Liability Act applies to “any claim for damages for harm resulting from negligence, regardless of whether the claim is brought in tort, in contract under statute or otherwise”: s 5A(1). Various terms used in this provision are defined in s 5: “ 5 Definitions

General principles are then stated by reference to “a risk of harm”: ss 5B, 5C, 5F-5I, 5L, 5M and 5R. None of these provisions expressly excludes the identification of harm as loss of an opportunity, but the inclusion of such a concept is at best awkward. For example, to speak in the terms of s 5B(2)(c), of “the burden of taking precautions to avoid the risk of harm”, is readily understandable in the context of harm as physical injury, but is at best obscure and semantically inappropriate, if the phrase “loss of an opportunity of a better outcome” is substituted for the word “harm”.
On the alternative approach, if provision of damages for the loss of a chance of a better outcome involves a new concept of causation, that construction would not be consonant with the principles relating to causation set out in ss 5D and 5E of the Civil Liability Act. In particular, s 5E requires that the plaintiff always bears the onus of proving, on the balance of probabilities, any fact relevant to the issue of causation.
Any change to the common law of negligence is a change to the common law (or general law) of Australia. Therefore, the form of one statute in one State may not be seen as adequate either for the analogical development of the common law (cf Esso Australia Resources Limited v Commissioner of Taxation of the Commonwealth of Australia [1999] HCA 67; 201 CLR 49 at [18]-[28] Brodie at [132]) or for restriction on any change (cf “paralysis” in Brodie at [132]).
The above provisions of the Civil Liability Act are substantially mirrored in a number of States and Territories: Civil Law (Wrongs) Act 2002 (ACT), ss 40, 43, 44, 45 (1) and (3) and 46; Civil Liability Act 2002 (Tas), ss 9, 11, 12, 13 and 14; Wrongs Act 1958 (Vic), ss 43, 44, 48, 49, 51 and 52; Civil Liability Act 2002 (WA), ss 3, 5B, 5C and 5D; Civil Liability Act 2003 (Qld), ss 9, 10, 11 and 12 and dictionary; and Civil Liability Act 1936 (SA), ss 3, 32, 34 (1) and (3) and 35.
Though this point was not the subject of argument on the appeal, it might be thought that it would be inappropriate for the general law to develop a concept of harm which departed from the assumptions underlying a tolerably uniform statutory definition of harm.
For the above reasons and as summarised below, we consider the approach adopted in Rufo and Gavalas to have involved a departure from conventional principles. There were no compelling reasons to adopt that approach and to do so was, in all the circumstances in our respectful view, plainly wrong, in the sense discussed above. In summary, we would not follow Rufo and Gavalas for the following reasons. First, the doctrine espoused formed part of no recognised stream of authority. Secondly, the doctrine espoused was not consistent with or permitted by Sellars. To the contrary, it fell outside the expression of principle as to commercial advantages and opportunities in Sellars, which otherwise stated orthodoxy of proof of causation. Thirdly, the setting of the law of torts on a new path of proof of causation, whether for medical negligence cases or torts generally, based on creation of risk and policy for fair recompense for loss is (and was in 2001 and 2004) a matter of high policy for the High Court. Fourthly, in neither Gavalas nor Rufo were the difficulties and complexities of the application of the doctrine considered. A discussion of the difficulties and complexities can be seen in the judgments in State of New South Wales v Burton (No 2) [2008] NSWCA 319. See also Halverson v Dobler [2006] NSWSC 1307 at [191]-[246] per McClellan CJ at CL (the question of loss of a chance not being dealt with on appeal: Dobler v Halverson [2007] NSWCA 335; 70 NSWLR 151). Other difficulties and complexities arise at the point of assessing the accrual of any cause of action for limitation purposes. Fifthly, no clear limitations have been formulated for the application of the new doctrine. No logical distinction limits its development to personal injury caused by medical negligence. Sixthly, the doctrine can be seen to be inconsistent with conventional authority, now reflected in the Civil Liability Acts, as to the nature of harm required to justify a finding of negligence. Seventhly, the general principle that a causal connection between the tortious conduct and the plaintiff’s injury must be established on the balance of probability is now expressly enacted in the Civil Liability Acts. Eighthly, there is no evidence that insurance companies or members of the public have adapted their personal or commercial relations in reliance upon these authorities. H. Claim for loss of benefits accruing on marriage
As a separate issue, albeit one which does not arise on the approach adopted above, the respondent made a claim for the loss of benefits accruing on marriage. The trial judge rejected this claim, for reasons discussed at [213]-[217]. In particular, his Honour referred to the decision of this Court in Rosniak v Government Insurance Office (1997) 41 NSWLR 608 at 617. There, Mason P, with whom Meagher JA and Clarke AJA agreed, stated that: “The proposition that loss of support from a marriage partner may be compensable where the defendant's wrong has deprived a plaintiff of that support is uncontroversial in principle.”
Mason P cited Warringah Shire Council v Jamieson (Court of Appeal, 19 December 1980, unreported) per Samuels JA at 35, who said that: “… [g]enerally, it will be appropriate to assume that a female plaintiff would have sought, and obtained, in marriage financial benefits roughly equal to those which she could have got in commercial or industrial employment.”
His Honour also referred to Government Insurance Office (NSW) v Mackie [1990] Aust Torts Reports ¶81-053 at 68,214, noting that case demonstrated that care needed to be taken to avoid double recovery when considering both loss of earnings and loss of support. Studdert J made particular reference to this at [213], when he observed that “care must be taken to avoid any element of ‘doubling up’”. Further, as Clarke AJA pointed out in Government Insurance Office (NSW) v Mackie (also referred to by Studdert J in the context of his reference to Rosniak), different cases needed to be dealt with differently. In that case, Clarke AJA had given, by way of an example, the case of a woman engaged to a wealthy man at the time of accident, in circumstances where there was evidence that the injured plaintiff had not planned to work after marriage. Clarke AJA said in that case, “the destruction of marriage prospects could not simply be swept aside”. His Honour commented that in other cases, the matter might be more appropriately weighed in the scales in determining the extent of any deduction for vicissitudes. This was the approach taken by Whealy J in Simpson v Diamond [2001] NSWSC 925.
His Honour distinguished the decision in Hines v The Commonwealth of Australia (1995) Aust Torts Reports 81-338, where Giles J (as his Honour then was) awarded a sum of $400,000 for loss of support on the basis that case involved an adult whose former fiancé was an accountant who had progressed in his profession by the time of trial. This case involved a young child, who was six years of age at the time she sustained her disabilities.
In those circumstances, his Honour considered it would not be appropriate to make a discrete allowance as separate head of damages for loss of marriage benefits. Rather, he reflected that loss in the allowance for vicissitudes in relation to the claim for the loss of earning capacity.
No error of principle has been demonstrated in the approach his Honour adopted. It accords with the authorities in this Court. It was an available approach to his Honour on the evidence. I. Orders

Gett v Tabet [2009] NSWCA 76

At a glance

Court

Decision date

Before

Source

Judgment

Parties

Legislation Cited (18)

Cases Cited (11)

Cited By (1)