King v Western Sydney Local Health Network

JUDGMENT 1BASTEN JA: The plaintiff, Tamara King, developed foetal varicella syndrome (FVS). The condition, which has caused severe physical and intellectual disability, resulted from her mother contracting varicella (or chickenpox) early in the second trimester of her pregnancy with the plaintiff. 2The source of the mother's exposure was the plaintiff's older sister. Immediately the parents realised the sister might have chickenpox, the mother sought advice from a doctor at Blacktown Hospital. The mother believed, and advised the doctor, that she herself had not suffered from chickenpox. She was therefore unlikely to have the necessary immunity. In accordance with standard medical practice in 2002, she should have been offered an intramuscular dose of 600 international units (IU) of varicella-zoster immunoglobulin (VZIG) to boost her defences to the virus. 3The plaintiff's mother was not offered such treatment and contracted chickenpox. It is not in dispute that the plaintiff's condition was a result of the mother's infection. 4The plaintiff sued the respondent, as the entity responsible for Blacktown Hospital and its medical staff, in negligence. The trial judge found that the legal duty of care owed by a medical practitioner to his or her patient extended to offering the mother VZIG. He also found that the mother would have accepted the treatment, if offered. He found that the treatment was not offered: King v Western Sydney Local Health Network [2011] NSWSC 1025. 5None of these findings was in dispute on this appeal. Rather, the issue in dispute was whether the trial judge was correct in concluding that the plaintiff had not established that, if the treatment had been administered, her mother would probably have avoided developing chickenpox. The background circumstances, the expert evidence and the reasoning of the trial judge in reaching this conclusion were lucidly set out in the judgment below and are equally lucidly explained by Hoeben JA. That matter need not be repeated. Whilst acknowledging the force of the reasoning, the conclusion reached is, in my view, erroneous, for the following reasons. 6First, the result is counter-intuitive and apparently anomalous. The legal duty of care required that the mother be offered a treatment which was available for the purpose of boosting immunity to the chickenpox virus. In breach of that duty, the treatment was not offered. The harm which was sought to be avoided came to pass. Normally that sequence would allow an inference as to the causal link between the breach and the harm. It is necessary to be satisfied that there is a sound reason why that inference was not drawn in the present case. 7The anomaly lies in the explanation of the duty, given by the trial judge at [271]: "In order to be regarded in the public health sense as being effective across a population, [the treatment] does not have to prevent infection in greater than one half of that population. In order to justify introduction by the publication of guidelines and subsequent prescribing practices, it may well be that the relevant authorities are satisfied if it is effective in a much lower percentage of cases. Matters of economics, and cost-benefit analysis are also relevant." 8It may readily be accepted that the existence of a tailor-made remedy, undoubtedly effective in some cases (eg young children), the absence of serious adverse effects and the likelihood of significant harm absent treatment, may in combination suffice to warrant a public health directive to offer the treatment to pregnant mothers, despite the effectiveness of the treatment not being scientifically proven. (If there were real doubts as to the effectiveness of the treatment, one might expect it to be accompanied by a warning, so as to allow the mother to consider alternative courses, such as termination.) This may be described as the "public health" approach. However, the legal duty which was upheld in the present case was a different matter. The duty pleaded, which was admitted by the defendant, included "an obligation to advise the plaintiff's mother of the availability of VZIG and its potential beneficial effects in preventing or ameliorating chickenpox": at [59]. The pleading also alleged that the Hospital "ought to have offered to administer the VZIG to the plaintiff's mother". 9The trial judge discussed the duty and content of the duty no further. He then turned to the application of s 5B of the Civil Liability Act 2002 (NSW). Although it is frequently said that s 5B relates to breach rather than duty, the two concepts are not entirely separate and independent. In any event, the breach identified by the trial judge, which reflected the content of the admitted duty, was that the Hospital "failed to act reasonably because they did not advise Mrs King about the availability of VZIG and its purpose and benefits, and they did not offer to administer it to her whilst she was a patient": at [95]. 10There is no hint in this unchallenged finding that the efficacy of VZIG in the circumstances in which the plaintiff's mother found herself was unproven. If it were, the legal duty should have included a warning to this effect, just as it should have included a warning if there were serious adverse side effects. As to the latter possibility, the trial judge expressly noted that no possible serious adverse effect was referred to in the evidence: at [93]. The trial judge should also have found that there was no evidence that the relevant authorities in Australia were satisfied that the drug was effective in a percentage of cases "much lower" than 50%, contrary to the suggestion at [271], set out above. The reason for making such a suggestion was, as will be further explained, the disconformity between the conclusion as to causation and the recommendations of medical authorities upon which the legal duty was founded. 11As explained by Hoeben JA, the reasoning of the trial judge with respect to causation focused squarely upon the results of a German study referred to as the "Enders Report". Yet the relevant figures in the Enders Report supported the plaintiff's case that VZIG was effective in 54% of cases and thus supported the view that, if administered, it would probably have prevented the plaintiff's mother contracting the illness and thus the plaintiff suffering the effects of FVS. 12However, the trial judge discounted the figures in the Enders Report on two bases. First, he noted that the dosage given to German mothers was almost triple the recommended Australian dose. After referring to certain evidence given by the experts, he concluded: "198 The result is that in applying the results of the studies recorded in the Enders Table to the Australian practice, I am satisfied that if only 600 international units were administered, then a higher number of the patients would have been infected with VZV than was shown in the Table. 199 It follows that the Table overstates the effectiveness of VZIG in VZV prevention when considering its application to the Australian administration practice." 13Nevertheless, he also concluded that it was "difficult, if not impossible, to quantify with precision" any adjustment to the table on this account: at [206]. 14The source of these findings requires further explanation. Enders and Miller had, in a publication in 2000, described the UK dose as having "a relatively low IgG antibody concentration of about 700 IU per adult dose" and had then stated that the "effect is to attenuate disease rather than to prevent infection", referring to a 1993 paper by Miller. That led to a question being put to the experts as to whether the effect of administering 600 IU would be "to attenuate disease caused by the varicella virus rather than to prevent infection": Joint Conference of Expert Witnesses, 18 March 2011, Q20. All of the witnesses answered somewhat cryptically, "could be both". In oral evidence, Dr Kesson said that the dose could have the effect of both attenuating disease and preventing infection but thought it "more likely to attenuate infection than prevent disease". Dr Hudson said it was "known that the higher dose is more likely to prevent infection and therefore attenuation of disease is more likely to be seen with a lower dose": Tcpt, 13/04/11, p 101(10)-(20). 15Whilst expressing those views, none of the experts suggested that the Australian practice involved a dose which was too low to be effective or that it would have been appropriate to give a higher dose. It will be necessary to return below to consider the significance of "attenuation" of the disease. It may be noted, however, that no expert suggested that the Enders Table was therefore inapplicable in assessment of Australian practice. 16The trial judge accepted these opinions as supported by further evidence, particularly from Dr Kesson, that the treatment provided antibodies to fight the virions and it was "just a numbers' game": at [191]. Although the trial judge set out a lengthy extract from Dr Kesson's evidence, at [192], in explanation of that phraseology, he did not explain that the answer was being given to a different question. She was commenting on an answer given in the report of the joint conference to question 21 which was: "On the balance of probabilities, does varicella-zoster immunoglobulin need to be administered to the patient as soon as possible for maximal effect in the patient." 17The focus of that discussion was the finding that VZIG appeared to be more effective when given 4-5 days after exposure, rather than in days 1-3. Dr Curtis hypothesised that where the virus was inhaled, viremia occurred "around 4 days later" and the antibody might be most effective if it arrived in the bloodstream at the same time as the virus: Tcpt, p 104(25)-(40). Whilst Dr Kesson was, of course, describing the same process, it should have been noted that her evidence was given in a different context. 18The trial judge also gained support from the comments in Enders and Miller, relying on the 1993 paper by Miller. However, these comments had been the basis for question 20, as noted by the Chair of the Joint Conference, and were thus the source of the experts' opinions, rather than independent support for them. 19The view that the Australian approved dose was less than adequate was inconsistent with what may be inferred to be the balance of informed medical opinion in this country in 2002, namely that 600 IU was a sufficient dose to be effective. Without evidence as to why that dose had been chosen in this country (a dosage of 625 IU had been accepted as the maximum appropriate dose in the USA and 700 IU as the appropriate dose in the UK), the results recorded in the Enders Report should not have been discounted on this basis. At the very least the exercise was inappropriate in the absence of evidence that the difference would be statistically significant and without reference to the significance of attenuation. 20A second element of discounting was supportable. However, the manner in which it was undertaken was not based upon statistical evidence. The trial judge adjusted the reported results to take into account evidence that at least 10% of exposed pregnant women would not have contracted chickenpox in any event. That was said to reduce the number for whom the treatment was effective from 54% to 44%: at [204]. 21Some form of adjustment was appropriate because, the plaintiff's mother having contracted the illness, it was clear that she was not in the 10% who would not be infected in any event. However, it is fallacious to say that VZIG is effective for only 44 out of every 100. Without VZIG, 90 in every 100 pregnant women are infected; with VZIG, the figure is 46 in every 100, a reduction in round terms of 50%. From the alternative perspective, of every 90 women at risk, 44 (approximately 50%) will not be infected if VZIG is given. 22Thirdly, the Enders Report appears to have been overvalued because it was "the best available" empirical evidence. That does not mean that it had high statistical accuracy. In fact, there was no evidence as to the statistical significance of a difference between say 48% and 52% or even between 46% and 54%. Further, despite the fact that it was described as a level 3 descriptive study (that is, it was not an experimental study involving randomised controls - see judgment below at [179]) it appears to have been treated as demonstrating the absence of a causal link, rather than as significant evidence in support of such a link, where no randomised trial was ethically possible. 23Fourthly, the question identified as decisive was the chance of the mother escaping infection. However, the ultimate question was the chance of the plaintiff escaping FVS. Empirical studies indicted that only 1%-2% of foetuses were seriously afflicted as a result of the mother's illness: judgment at [75]. One might have expected a greater focus on why at least 98% of foetuses escape FVS. To the lay mind, the mother's viral load might be a possible factor. If so, that would invite attention to the expert evidence that VZIG might, even in cases where disease was not prevented, have resulted in an attenuated form of the disease. On the basis that prevention occurred in about 50% of cases, even a slight reduction in infection of foetuses resulting from an attenuated form of the disease, would be sufficient to establish causation on the balance of probabilities. 24As noted above, the joint experts had been expressly asked about attenuation or prevention in question 20, and the trial judge expressly relied upon evidence suggesting that a lower dose might have a greater effect in attenuating disease than in preventing infection. The issue was squarely raised in the report of Professor Nigel Curtis of 17 March 2008. Dr Curtis was Professor of Paediatric Infectious Diseases at the University of Melbourne, Head of Infectious Diseases Unit, Royal Children's Hospital Melbourne and Leader of Microbiology and Infectious Diseases Research Group, Murdoch Children's Research Institute. In respect of developing foetal varicella syndrome where chickenpox was contracted by the mother during the early stages of pregnancy, Professor Curtis stated (Report, p 2): "The most reliable data comes from the largest study (Enders et al 1994), which included a total of 1,373 women. This study reported a risk of 0.4% (95% confidence intervals (CI) 0.05-1.5%) in the first twelve weeks of pregnancy and a risk of 2% (95% CI 0.8%-4.1%) from 13 to 20 weeks gestation." 25In addressing a question as to the chance that the proper administration of VZIG would have either eliminated the injury to the plaintiff or reduced its effect, Professor Curtis stated (Report, p 7): "I believe there is strong indirect evidence to support the contention that [VZIG], when given to a pregnant [woman] after exposure to chickenpox, does have a role in preventing FVS or ameliorating its effects: 1. Given that [VZIG] prevents the development of chickenpox in a large proportion of exposed individuals, including women during pregnancy, it is a logical step to conclude that it may therefore prevent FVS. If the pregnant mother does not develop a viremia with the varicella-zoster virus, it is reasonable to conclude that the baby cannot be infected and subsequently develop FVS. 2. There is some evidence in the published literature to suggest that [VZIG] does in fact prevent FVS .... In the study published by Enders et al 1994, nine babies were reported with features suggestive of FVS and none of their mothers had received [VZIG] following exposure to chickenpox during pregnancy. In contrast, of 97 other women who had received [VZIG] following chickenpox exposure during pregnancy (but who nonetheless developed clinical chickenpox) none had a baby with FVS. 3. In many other contexts, even when [VZIG] does not completely prevent chickenpox or its consequences, it does ameliorate its effect, leading to disease of lesser severity." 26In table 4 to his report, Professor Curtis summarised the Enders 1994 results in this way: "Study group - 1,373 women with chickenpox during pregnancy (+ 366 with herpes-zoster) Key results - 9 women had a baby with features suggestive of FVS - none of these mothers had received [VZIG] - 97 women who received VZIG developed chickenpox - none had a baby with FVS" 27In oral testimony, Professor Curtis repeated and explained the summary of those results, noting that "there is some indirect evidence that the babies born to mothers given VZIG were less likely to be infected": Tcpt, p 121(13). 28The trial judge dealt with this evidence, which was published in The Lancet Vol 343, pp1548-1551 (18 June 1994), including the conclusion, noted by the trial judge at [250] that "[p]ost-exposure prophylaxis with [VZIG] attenuates disease and there was evidence from the present study that it may also reduce the risk of foetal infection". The trial judge did not, however, refer to the statement which was part of the material relied upon by Professor Curtis at 1550: "No cases of congenital varicella syndrome or zoster in infancy occurred among the 97 pregnancies in which maternal varicella infection followed post-exposure anti-varicella-zoster varicella-zoster immune-globulin prophylaxis, although specific IgM antibody was found in one in 89 (1.1%) cord bloods tested." 29The trial judge nevertheless concluded that the paper "does not reveal whether the cohort of pregnant women were administered VZIG at any stage ... had it included universal administration of VZIG within a short time after exposure, the size of the cohort ought to have provided compelling scientific results ... it did not": at [251]. He then continued at [252]: "As Professor Curtis reports, there were 97 cases within the cohort in which maternal varicella infection followed administration of post-exposure VZIG. But it is not possible to draw meaningful conclusions from this statement alone because the study does not reveal how many of the cohort received VZIG." 30This statement appears to miss the point which Professor Curtis was deriving from the study, namely that of the nine babies who may have had FVS, none of their mothers had received VZIG, whereas of the 97 women who received VZIG but nonetheless developed clinical chickenpox, none had a baby with FVS. The omission in the reasoning of the trial judge appears to have resulted from a concentration on whether VZIG would probably have prevented the mother contracting chickenpox, rather than the ultimate question, namely whether VZIG would have probably prevented the plaintiff developing FVS. 31It follows that the plaintiff's complaints that the trial judge adopted an overly mechanistic approach, placing too much weight on scientific evidence which, properly understood, either supported or was neutral as to probable causation, has been made good. The question is whether, despite those errors, the trial judge was right to dismiss the causal link as unproven. In my view he was in error. The scientific evidence outlined above demonstrated that a causal connection was more probable than not. But even if it did not, that did not determine the legal question. As explained by Spigelman CJ in Seltsam Pty Ltd v McGuiness [2000] NSWCA 29; 49 NSWLR 262 at [137]: "In Australian law, the test of actual persuasion does not require epidemiological studies to reach the level of a Relative Risk of 2.0, even where that is the only evidence available to a court. Nevertheless, the closer the ratio approaches 2.0, the greater the significance that can be attached to the studies for the purposes of drawing an inference of causation in an individual case. The "strands in the cable" must be capable of bearing the weight of the ultimate inference." 32There was other material to support the inference. As the trial judge noted, not only did Guideline 13 issued by the Royal College of Obstetricians and Gynaecologists (July 2001) state in unequivocal terms that a pregnant woman who is not immune to chickenpox should be given VZIG (judgment at [132]), but standard texts, including the Australian Immunisation Handbook (7th ed, March 2000), made a similar recommendation, also without qualification or equivocation, as did a publication entitled Management of Perinatal Infections (2002, Australasian Society for Infectious Diseases): judgment at [90] and [91] respectively. This material was more than sufficient to justify a finding of causation which, in terms of the duty identified, should be made. 33Finally, the factors noted at [8] above warranted recovery. Were it otherwise, the legal duty which was upheld and breached would be an empty gesture. The very risk which the treatment was designed to alleviate, and which eventuated, was, on the approach adopted below, not shown to be caused (and could not be shown to be caused on the available empirical studies) by the breach of duty. If that were correct, a real question would arise as to whether a legal duty of the kind formulated was appropriate. However, the existence and content of the duty pleaded were not challenged. 34Upon this reasoning, the present case falls within s 5D(1)(a) of the Civil Liability Act, "factual causation" having been established. In the alternative, it might be thought that a case involving the elements to which reference has been made, but lacking a sufficient empirical basis for reaching a conclusion about the effectiveness of a recommended treatment, fell within s 5D(2) as an exceptional case of the kind illustrated by Bonnington Castings Ltd v Wardlaw [1956] AC 613, referred to in Strong v Woolworths Ltd [2012] HCA 5; 86 ALJR 267 at [24]-[27]. At trial, it was said that the plaintiff eschewed reliance on s 5D(2). However, like the relationship between duty, content and breach, questions of factual causation and scope of liability, as separately identified in s 5D, do not readily fall into separate and independent watertight compartments. Valuable as it is to separate the "factual" and "policy" elements of causation, the separation is, to an extent, an artefact. It would be a triumph of form over substance to deny the plaintiff recovery on that basis. 35I would allow the appeal, set aside the orders of the trial judge and give judgment for the plaintiff for damages to be assessed. The respondent must pay the appellant's costs in this Court and the costs of the trial of the separate question of liability. 36HOEBEN JA: Nature of appeal The appellant seeks leave to appeal from the finding by the primary judge on 7 September 2011 in favour of the respondent on the issue of liability. The application for leave and the appeal were heard together. 37It was the appellant's case at trial that she developed Congenital Varicella Syndrome (CVS) as a result of her mother becoming infected with varicella (chickenpox) while she was pregnant with her. The appellant claimed that when her mother attended the Blacktown Hospital on 6 May 2002, she ought to have been given an injection of Varicella Zoster Immunoglobulin (VZIG) and that if she had, this injection would, on the probabilities, have prevented her mother being infected with varicella and as a result, she would not have developed CVS. 38The primary judge found in favour of the appellant on the issue of duty of care and breach of duty, but found against her on the issue of causation. Accordingly, he entered judgment in favour of the respondent. Factual background 39The following definitions are used throughout this judgment and their meaning is not controversial: (a) Varicella (chickenpox) is a highly contagious disease caused by the varicella-zoster virus. The disease is characterised by fever, malaise and an itchy rash. (b) Varicella-Zoster Virus (VZV) is a virus of the herpes family that is transmitted by respiratory droplets or direct personal contact. (c) Congenital Varicella Syndrome (CVS) which is caused by chickenpox infection in a pregnant woman, may be characterised by any one or more of the following: (i) Skin scarring. (ii) Eye defects. (iii) Hypoplasia of the limbs. (iv) Neurological abnormalities, including microcephaly, cortical atrophy, mental retardation and dysfunction of the bowel. (d) Varicella-Zoster Immuno Globulin (VZIG) is a sterile solution of the globulin fraction of human plasma which is administered by injection intra-muscularly. In the evidence and medical literature, Zoster Immuno Globulin (ZIG) is also referred to. Unless specifically noted, VZIG and ZIG refer to similar products and can be regarded as interchangeable. (e) Significant exposure to VZV infection is defined to include living in the same household as a person with active chickenpox or herpes zoster (shingles), play contact of longer than one hour, classroom contact or other close, prolonged exposure. 40The factual findings made by the primary judge are not in issue. The following summary is taken from those findings. 41On the evening of Sunday 5 May 2002, the appellant's father first noticed a pink rash on the left arm of his daughter Shania (the appellant's older sister). The appellant's mother (Mrs King) saw the same rash on that evening. At the time, Shania slept in the same bedroom as her mother. 42Early on the morning of 6 May 2002, Shania was diagnosed by the family general practitioner as having chickenpox. He advised that the appellant's mother should sleep in a separate bedroom from Shania. The reason for that advice was that the appellant's mother was pregnant with the appellant. Her estimated date of confinement was 9 November 2002. 43At about 10.15am on 6 May 2002, the appellant's mother presented at the Blacktown Hospital. She did so because she was having vaginal bleeding and back and abdominal pain. At about 11.45am she was attended by Dr Davidson, who noted by reference to the hospital notes, that she was in the fourteenth week of her pregnancy. Dr Davidson noted "daughter developed chickenpox yesterday - she does not believe she has had chickenpox". 44In relation to the issue of chickenpox, Dr Davidson noted: "D/W O+G re need for zoster immunoglobulin - is in second trimester therefore not given - Zoster IgM IgA - he will check at ANC on 9/5". 45His Honour interpreted that note to mean that Dr Davidson had discussed with an obstetrician and gynaecologist, the need for the appellant's mother to be given VZIG and that the obstetrician and gynaecologist had advised that since she was in her second trimester, there was no need for the administration of the VZIG on that day. Dr Davidson advised that a blood test ought to be taken to check for the presence of antibodies to chickenpox, which was intended to establish whether the patient was immune to VZV. Dr Davidson was advised that the results would be checked at the ante-natal clinic on 9 May 2002 and such treatment as was then appropriate would be provided. 46It was common ground that the appellant's mother was not treated at that time but was reassured, blood was taken and she was told that a obstetrician and gynaecologist would review her on 9 May 2002 at the ante-natal clinic. She left the hospital at about 12.45 pm on 6 May 2002 and underwent an ultrasound examination at the rooms of North West Radiology. Nothing turns on the result of that examination. 47On 9 May 2002 the appellant's mother attended the ante-natal clinic at Blacktown Hospital. Nothing of significance appears to have occurred on that occasion. In particular, there does not appear to have been any consideration or discussion of the issue of her exposure to Shania's chickenpox on that occasion. 48On 17 May 2002 the appellant's mother was diagnosed as suffering from chickenpox "since yesterday". On 19 May she attended Blacktown Hospital Emergency Department as a result of a rash on her chest for a couple of days and all over her body, having started "last Friday" (i.e., 17 May 2002). She remained in hospital until 22 May 2002. His Honour concluded that the symptoms of chickenpox had commenced on 17 May 2002. 49On 31 October 2002 she was admitted to the hospital in preparation for the appellant's birth, which occurred on 1 November 2002 at about 9pm. By 6 November 2002 it was suspected that the appellant had a congenital varicella infection. Dr Cheryl Jones, a paediatric infection specialist, reported on 6 December 2002 that the appellant was likely to have CVS. The primary judge's reasons 50The primary judge found that the content of the duty of care owed by the respondent to the appellant (which was not disputed) was "an obligation to advise the appellant's mother of the availability of VZIG and its potential beneficial effects in preventing or ameliorating chickenpox and an obligation to have offered to administer VZIG to her". 51It was common ground that the Civil Liability Act 2002 (CLA) applied to the proceedings and that breach of duty of care was to be decided in accordance with s5B. By reference to that section, his Honour found that the relevant "risk of harm" was the risk of the appellant being born with CVS. 52His Honour found that Dr Davidson knew of the "risk of harm" which was why she had consulted with an obstetrician and gynaecologist on 6 May. In the alternative, his Honour found that it was widely known that there was such a risk of CVS in children born of mothers who were exposed during pregnancy to VZV (1 - 2 percent) and that the Hospital and its medical staff ought to have known of that risk. His Honour found that s5B(1)(a) was therefore satisfied. 53There was no issue between the parties that the risk of harm was not insignificant (s5B(1)(b)). Having regard to the known rate of occurrence of CVS in children born of mothers exposed to VZV and the potential for serious consequences for those children, his Honour found that the requirement had been satisfied. 54In relation to "reasonable precautions" (s5B(1)(c) and s5B(2)), his Honour noted that the features of CVS were commonly permanent and could be very disabling and that the possibility of their occurrence could not be ignored. The administration of VZIG was directed to preventing or ameliorating maternal chickenpox with a possible result that the infection of the foetus would be prevented. In those circumstances, his Honour concluded that the risk of harm was one against which a reasonable person would take precautions. 55His Honour noted that in the Joint Expert Report (exhibit C), five of the six experts who gave evidence concurrently (Dr Jones offered no comment) agreed that: (a) On 6 May 2002 the appellant's mother ought to have been advised of the risk for her and [the appellant] of developing chickenpox or CVS as a result of Mrs King's exposure to Shania's infection; (b) On 6 May 2002 the appellant's mother ought to have been advised of the possibility of treatment by VZIG; (c) In Australia in 2002 it would not have accorded with proper and reasonable professional practice, to have failed to administer VZIG to a pregnant woman where exposure had occurred within a 96 hour period prior to administration. 56The primary judge concluded that "a reasonable person in the position of the hospital, acting through its medical staff, would have taken the precaution which the plaintiff alleged should have been taken, namely, that Mrs King ought to have been advised about the availability and potential benefits of VZIG, and the appropriateness in her case of her being administered that preparation" (Red 254 K - M). 57His Honour had regard to s5O CLA. His Honour concluded that the respondent had not made out a defence under that section. In particular, the respondent had failed to establish that the appellant's mother was infected with VZV on either 1 or 2 May 2002 so that the infection occurred outside the 96 hours from the time of first exposure to the virus. His Honour found that the treatment provided by the respondent did not accord with Guideline 13 of the Royal College of Obstetricians and Gynaecologists. There was no challenge, by way of Notice of Contention, to those findings in the appeal. 58His Honour next considered the issue of causation. It was common ground that ss5D and 5E CLA applied. 59Sections 5D and 5E are relevantly in the following terms: "5D(1) A determination that negligence caused particular harm comprises the following elements: (a) that the negligence was a necessary condition of the occurrence of the harm (factual causation), and (b) that it is appropriate for the scope of the negligent person's liability to extend to the harm so caused (scope of liability). (2) In determining in an exceptional case, in accordance with established principles, whether negligence that cannot be established as a necessary condition of the occurrence of harm should be accepted as establishing factual causation, the court is to consider (amongst other relevant things) whether or not and why responsibility for the harm should be imposed on the negligent party. ... (4) For the purpose of determining the scope of liability, the court is to consider (amongst other relevant things) whether or not and why responsibility for the harm should be imposed on the negligent party. 5E In determining liability for negligence, the plaintiff always bears the onus of proving, on the balance of probabilities, any fact relevant to the issue of causation." 60At trial, senior counsel for the appellant eschewed reliance upon s5D(2) and did not submit that the case was an exceptional one to which those provisions applied. As a result, his Honour did not consider 5D(2) and no submissions were made in relation to it. His Honour noted that neither party submitted that it was not appropriate for the scope of the defendant's liability, if established, to extend to the plaintiff's CVS. Section 5D(1)(b) was not in issue and his Honour did not have regard to it. 61His Honour concluded that the issue of causation was encompassed by the term "factual causation", as used in s5D(1)(a). By reference to the "but for" test, his Honour articulated the factual causation issue as: "Had, on 6 May 2002, the medical staff at Blacktown Hospital, appropriately advised about and administered 600 international units of VZIG to Mrs King, the plaintiff's mother, she would, on the balance of probabilities, not have contracted varicella on 16 May 2002." (Red 268 U - W) 62There was a preliminary question. His Honour determined it in favour of the appellant. The question was whether if so advised, and if VZIG had been offered to her, the appellant's mother would have accepted the administration of VZIG. No challenge to that finding was made by the respondent in the appeal. 63In order to establish causation, the appellant relied upon scientific and expert evidence. The scientific evidence comprised studies which were to be found in medical and scientific literature. The expert evidence was provided by a panel of six experts, who gave evidence concurrently. Before doing so, they prepared a joint report (exhibit "C"). The experts who gave evidence were: Professor Curtis from the Royal Children's Hospital Melbourne - paediatric infectious diseases. Associate Professor Kesson from the Children's Hospital Westmead - paediatric infectious diseases. Dr Hudson from the Royal North Shore Hospital - adult infectious diseases. Professor Trudinger, head of the Maternal Foetal Medicine Service at Westmead Hospital - obstetrician and gynaecologist. Professor Gillian Turner - clinical geneticist. Dr Robert Jones - paediatric neurologist. 64The part played by Dr Jones was limited. His expression of opinion was limited to whether the appellant suffered from CVS. 65His Honour recorded the response of the experts to the factual causation question at [61] hereof, as follows: (a) Professor Curtis: "on the balance of probabilities [the chickenpox] would have been prevented ..." (Black 150E) (b) Professor Turner: "50 percent chance that [the administration of VZIG would] stop [the patient] getting varicella ..." (Black 150R) (c) Professor Trudinger: "uncertain what benefit there was" (Black152J) (d) Associate Professor Kesson: "less than 50 percent [and as time progresses, effectiveness] becomes less and less and less ... until it would have no effect" (Black 152P) (e) Dr Hudson: "my answer to, is it more likely than not, is 'no'" (Black 153K) His Honour concluded that only Professor Curtis was prepared to advance the opinion that it was more likely than not that if VZIG had been administered, the appellant's mother would not have developed varicella. On the basis that the appellant's case depended on the opinion of Dr Curtis being accepted, his Honour examined his evidence to see whether this opinion ought be accepted. 66The following is his Honour's analysis of the evidence of Professor Curtis. 67In reaching his opinion of Professor Curtis his Honour relied upon a combination of the following: (a) The statistics contained in the table which was reproduced in answer to Question 21 in the joint report (Exhibit C) which in turn was derived from Table 16.7 in Enders and Millers chapter "Varicella and herpes zoster in pregnancy and the newborn" in Arvin et al (2000) (Ex L) (Blue 255). This was referred to as the "Enders Table". (b) Studies to which he referred in Table 3 of his report of 17 March 2008 (Ex D) (Blue 583); (c) Studies to which he referred in Table 4 of his report of 17 March 2008 (Ex D) (Blue 584); (d) His own clinical experience. 68Professor Curtis relied upon the Enders Table to demonstrate that: "VZIG prevented infection in 54% of women given VZIG 1-2-3 days after exposure - therefore on the balance of probabilities the plaintiff's mother would not have been infected with chickenpox. (Ex C, footnote p32) (Blue 709W)." Professor Curtis regarded the figures in the Enders Table as conservative. 69Because the Enders Table played such an important part in the reasoning of Professor Curtis, the primary judge examined its reliability. The Enders Table was contained within a publication which the experts accepted was peer-reviewed by authors regarded as experts in the field. 70The other experts, however, qualified their acceptance of the Table as follows: (a) It was not compiled from the results of a randomised controlled clinical trial, but was derived from one or more observational studies ("the study quality issue"). (b) The dosage used for the patients in the Table was 1750 - 2000 international units of VZIG, whereas the standard dose used in Australia in 2002 was 600 international units which was the dose which ought to have been administered to the appellant's mother ("the differential dosage issue"); and (c) The Table did not, in reproducing the numbers of patients who avoided being infected with varicella, take into account those who would have failed to become infected, even if they did not receive VZIG ("the virus strike rate issue"). 71In relation to the "study quality issue", his Honour took into account the July 2001 Guidelines on "Chickenpox in Pregnancy" of the Royal College of Obstetricians and Gynaecologists (Exhibit "L") (Blue 72H) which classified the level of evidence in scientific studies: "Classification of Evidence Levels Ia Evidence obtained from Meta-analysis of randomised control trials. Ib Evidence obtained from at least one randomised control trial. IIa Evidence obtained from at least one well-designed control study without randomisation. IIb Evidence obtained from at least one other type of well-designed quasi-experimental study. III Evidence obtained from well-designed non-experimental descriptive studies, such as comparative studies, correlation studies and case studies. IV Evidence obtained from expert committee reports or opinions and/or clinical experience of respected authority." 72It was accepted by the experts, including Professor Curtis, that the Enders Table was in essence an observational study. It was not a randomised control trial. It therefore fell into the category of level 111 evidence. The experts agreed that if other studies of a higher evidentiary level were available, they would form a much sounder basis for drawing scientific conclusions. However, notwithstanding this characterisation of the study, and the fact that 212 patients was a relatively small number for statistical satisfaction, the Enders Table was the best available material from which to understand the effects of VZIG in pregnant women. 73The "differential dosage issue" arose because of the different practices around the world about the dose of VZIG administered to non-immune pregnant women, who had been exposed to VZV, as a matter of routine. The routine dosage in Germany in the patients included in the Enders Table was 1750 - 2000 international units. The recommended and routine dosage administered to patients in Australia was 600 international units in 2002. In the United Kingdom, a typical dose was 700 international units. 74The issue which arose in the course of the concurrent evidence was what, if any, allowance ought to be made to the statistical results in the Enders Table and in the interpretation of it, to account for this significant difference in dosage practices when applying the results to Australia in 2002. 75The primary judge reviewed the evidence on that issue: Associate Professor Kesson: "The effect of [VZIG] at 600 international units per adult dose, could be both to attenuate disease and to prevent infection. However, I think it is more likely to attenuate infection than prevent disease". (Black 108G) She used the expression "attenuate" to mean to reduce or lessen, the adverse effects of the disease. Associate Professor Kesson explained that the work which VZIG did as an antibody to counteract the VZV was "just a numbers game" (Black 111G). She said: "But I would say from a pathophysiological point of view that what you are trying to do with antibody is prevent establishment of infection and what you do with cellular immunity is [to] clear established infection. They are different arms of the immune response. So here we are trying to prevent establishment of infection. When one initially gets exposed to varicella, there will be a fixed amount of virus particles or virions, to which one is exposed. And if you have antibody, there is a fixed amount of antibody which you either had in your body, because you are immune, or that we administer to you as a form of passive immunity. And this is basically a numbers game, and if there is not sufficient antibody to neutralise, which is an immunological concept to prevent establishment of infection, to neutralise the virus, then the virus will continue to replicate. So as time goes on, from the point of exposure in your body and its entering cells and replicating you're generating more and more virus particles. So it is just a numbers game. This is why antibody has no therapeutic effect when infection is established, because there are so many virus particles, compared to the antibodies." (Black 111D-K) 76Dr Hudson said: "It is known that the higher dose is more likely to prevent infection and therefore attenuation of the disease is more likely to be seen with a lower dose than with a higher dose." (Black 108J) 77Professor Curtis did not disagree with that proposition but his acceptance of it was qualified: "I think the dose issue is slightly overblown. I am not aware of head to head studies of different doses. There have been different doses, I accept that. It is difficult to compare different doses and different brands of antibodies, because we're interested in both the quantity and the functional aspects of the antibodies. So I have mentioned, I am not aware of any head to head studies that show one concentration is higher than another. Intuitively, it would make sense that a higher dose would be better than a low dose. One accepts that. But that is not to say there is a threshold over which it is sufficient, so 600 is enough and 700 is still enough." (Black 153T-W) 78The other experts did not proffer any contradictory opinion to that expressed by Associate Professor Kesson and Dr Hudson. Accordingly, the primary judge accepted that evidence. In doing so, he noted that the evidence was supported by the conclusions and remarks of Dr Enders in her publication of which the Table formed a part. At Blue 255K-R Dr Enders discussed the difference between the routine dose in the United Kingdom of 700 international units and that in Germany which was between 1750 and 2000 international units. 79In relation to the United Kingdom, Dr Enders said: "Their effect is to attenuate disease rather than prevent infection". Of the German experience, Dr Enders referred to the Table and said: "According to our experience in 212 seronegative pregnancies, about half of the mothers have clinical varicella virus if VZIG is given in the recommended dosage ... and a further 5% have subclinical infection." 80The primary judge set out his conclusions in relation to the "differential dosage issue" as follows: "196 As I see it, the end point of the differential dosage issue is that when considering the application of the Enders Table to Australian practice, where a much smaller dosage is used, it is likely that some of the people who are recorded in the Table as showing no infection, would, in the Australian context, have been likely to have been infected with varicella because the lower dose would not have prevented the infection, although it may have attenuated it. 197 It is not possible to now be precise about how many people that would have been. Accordingly, it is not possible to adjust with accuracy the actual numbers in the Enders Table to account for this issue. However, it is clear that with a lower dosage of VZIG a larger number of patients in any identified cohort are likely to be infected, than those in the cohort studied in the Enders Table. 198 The result is that in applying the results of the studies recorded in the Enders Table to the Australian practice, I am satisfied that if only 600 international units were administered, then a higher number of the patients would have been infected with VZV than was shown in the Table. 199 It follows that the Table overstates the effectiveness of VZIG in VZV prevention when considering its application to the Australian administration practice." (Red 278 I-X) 81The primary judge had regard to the "virus strike rate issue" because it was common ground among the experts that where there had been household contact, statistics over time demonstrated that 85 - 90 percent of persons who were not immune to VZV, were likely to become infected. In other words, 10 - 15 percent of those exposed and non-immune persons would not be infected. This meant that when considering the Enders Table and what conclusions could be drawn from it, it was necessary to take into account that of the total group of 212 exposed non-immune persons to whom VZIG was administered, 10 to 15 percent would not have been infected in any event, even if VZIG had not been administered. 82As a result, his Honour concluded that the Enders Table needed to be adjusted so as to take account of the consequences of the "virus strike rate issue". Using a figure of 10 percent, his Honour concluded that of the 212 group members, 21 would not have been infected with varicella regardless of whether VZIG was administered. 83Professor Curtis agreed that it was appropriate to adjust the Enders Table to take account of that issue, but suggested that the only adjustment which needed to be made was to the "non-infected" group on the Table (Black 168P, 180U). His Honour considered that Professor Curtis was mistaken in that evidence and noted that the issue only arose for the first time during the course of the concurrent evidence and that Professor Curtis "had only a brief time to consider the real impact" of the issue. When his Honour adjusted the Enders Table for the virus strike rate, the percentage of those infected became 46 percent and those not infected, 44 percent. 84His Honour set out his conclusions, based on the Enders Table, as follows: "206 Without making any adjustments because of the differential dosage issue which are difficult, if not impossible, to quantify with precision, but merely to make the adjustments, to which all experts agreed in principle by reason of the virus strike rate issue, the Enders Table, as the best available scientific evidence, demonstrates that infection was prevented by the administration of VZIG in fewer than one half of the studied cohort. Put another way, notwithstanding the administration of VZIG, over one half of the studied cohort were infected with chickenpox. 207 Thus I conclude, based on this Table alone, that it has not been proved, more probably than not, across a population, that if VZIG had been administered it would have prevented infection. The Table does, however, clearly demonstrate that it was possible that infection may have been prevented. 208 It is important to note that even if I had concluded that the Enders Table had demonstrated that more patients in the cohort had not been infected if they had received VZIG than if they had not received it, whilst ever the numbers who did get infected, were at a level higher than an insignificant one, that would not have been sufficient to draw the conclusion which Professor Curtis did. It is important to note that even if I had concluded that the Enders Table supported the finding that it was probable that VZIG administration prevented chickenpox across a population, that would not have been sufficient of itself to draw the conclusion which Professor Curtis did that it was probable that VZIG would have prevented Mrs King from contracting chickenpox." (Red 281 C - T) 85His Honour's reasoning in reaching that conclusion was as follows: The essential question to be answered was "what would have happened to Mrs King as an identified individual and not just what would have happened across a population?" That question was not to be answered simply by translating the scientific results to the legal test of causation, e.g., if on the unadjusted Enders Table, 54 percent were found not to have any varicella infection. General causation which was the only conclusion open on the Enders Table, whether adjusted or not, was only evidence of possibility and not probability (Seltsam Pty Limited v McGuiness; James Hardie & Coy Pty Limited v McGuiness [2000] NSWCA 29; 49 NSWLR 262 at [60], [78] per Spigelman CJ (Davies AJA agreeing). 86His Honour then examined the other studies referred to in Tables 3 and 4 of the report of Professor Curtis of 17 March 2008 (Exhibit D). His Honour identified a number of difficulties with those studies. In two cases, the articles had not been tendered so his Honour could do little more than note their subject matter. In other cases, VZIG had been administered in quite different contexts and they were studies which examined both prevention and attenuation without distinguishing between the two. It is not necessary to set out the detailed analysis which his Honour conducted of those studies which were available to him. It is sufficient to set out his conclusions. 87With respect to the Table 3 studies, his Honour said: "234 The studies identified in Table 3 of Professor Curtis' report can be accepted as demonstrating that: (a) in the case of peri-natal chickenpox (that is, 27 days before to 28 days after birth), where VZIG was administered to the newborn, between 50 per cent and 61 per cent developed infection but the infection was, on the probabilities, attenuated; (b) in the case of older children (ie, aged between 21/2 years and 6 years), ZIG was successful in preventing varicella in all six cases in which it was administered; (c) in the case of immunocompromised patients, VZIG was successful in preventing varicella in 80 per cent of the patients. 235 These studies do not, and Professor Curtis does not himself, suggest that the results are capable of direct application to the segment of the population in which Mrs King fell (that is, non-immune pregnant women). Nor do they suggest direct application to the Australian experience, because of the potential for differential dosages. 236 The evidence does not identify any particular reason for the disparity in results between the newborn cohort with perinatal exposure and infection, and the other two cohorts. 237 I am not able to find in any of the expert evidence and other material dealing with these studies any explanation for the difference in the study results, save to note that it seems clear that VZIG works differentially in different groups of patients to whom it is administered. These studies, by themselves, do not compel a conclusion, on the balance of probabilities that in the relevant cohort, that is, non-immune pregnant woman, VZIG will prevent varicella infection. They do provide some additional material about the effectiveness of VZIG in different cohorts." (Red 286T - 287P) 88With respect to the Table 4 studies, his Honour said: "253 From these studies, referred to in Table 4 of Professor Curtis' report, I am unable to draw any conclusion which assists in the resolution of the issue posed in this case. 254 I do not accept that they provide any support for, nor do they detract from, the opinions expressed by Professor Curtis. However, they do add to a limited extent, to the general knowledge base of VZIG effectiveness." (Red 290P - T) 89In forming his opinion, Professor Curtis had also relied upon his personal observations of the effectiveness of VZIG in the course of his clinical practice where it was highly effective to prevent varicella infection. He said: "In my everyday practice we give VZIG to children who are severely immunocompromised, children with cancer who have no immunity whatsoever. It is a highly effective product. It is considered a medical emergency to treat those patients, and I have never seen VZIG fail, in those circumstances. So I have a much higher opinion of VZIG, based on my personal observation and on the literature. I see that Table as providing conservative estimates." (Black 158R - T) 90Professor Kesson, who had a clinical practice similar to that of Professor Curtis, agreed with his opinion in relation to immunocompromised children. However, she was of the opinion: "I don't think that (this clinical practice) is analogous to this situation of giving VZIG to a pregnant woman to protect the foetus, because we have no data whatsoever that it is efficacious." (Black 162T - U) 91His Honour posed this question to the experts: "HIS HONOUR: There is a question I have been meaning to ask the panel. In the course of the evidence and the joint report, reference has been made to immunocompromised patients and the effect which VZIG may have in patients of that kind. And there have been references to pregnant women. Are pregnant women necessarily immunocompromised by reason of their pregnancy only? WITNESS KESSON: Strictly speaking yes but in a practical sense in their vulnerability to varicella no. ... WITNESS TRUDINGER: Pregnancy is a unique situation, of course, because the mother tolerates a graft of the foetal material which is tissue foreign to her. HIS HONOUR: Partially foreign to her. WITNESS TRUDINGER: Partially foreign to her and some of them are totally foreign to her. But there are mechanisms that operate on the immune system to allow that to happen. So that, when you look at those specific mechanisms, they may be classed as immunocompromised with respect to those mechanisms but they can mount a normal immune response away from that. HIS HONOUR: To viral disease of some other sort? WITNESS TRUDINGER: Yes. HIS HONOUR: And is that what you are getting at Dr Kesson? WITNESS KESSON: Yes. HIS HONOUR: Dr Curtis do you have a view about that? WITNESS CURTIS: I agree it is a difficult question because of the definition of immuno-compromised and what different people actually mean by that. I think people often talk about pregnancy being an immunocompromised state. It is semantic I guess." (Black 178W - 179P) 92In relation to the clinical experience of Professor Curtis, his Honour found: "262 Whilst I accept that Professor Curtis and Associate Professor Kesson in the course of their clinical practice of treating severely immunocompromised children have found VZIG to be highly effective in preventing or ameliorating varicella infection, I am not satisfied that it is a sufficient basis for concluding that VZIG is more likely than not to be effective in preventing varicella infection in pregnant women, generally, or Mrs King in particular." (Red 292F - J) 93His Honour summarised his conclusions in relation to the opinion and evidence of Professor Curtis as follows: "263 I do not accept Professor Curtis' view that it is more likely than not that had VZIG been administered to Mrs King on 6 May 2002 she would not have contracted chickenpox. 264 Although the reasons for this appear from the lengthy discussion which precedes this part of the judgment, it is perhaps best to repeat them, in short form: (a) the Enders Table, when adjusted properly to reflect the virus strike rate issue, in fact demonstrates that VZIG did not prevent infection in over half of the studied cohort; (b) other adjustments ought properly to be made to the results of that Table which cannot now be precisely quantified, which would have the effect of making the prospect of VZIG preventing infection, if administered in Australia, less likely than the results recorded in the Table; (c) other studies, properly considered, do not support Professor Curtis' opinion; (d) the observation in clinical practice of the effectiveness of VZIG in immunocompromised children does not assist in elucidating the question about the effectiveness of VZIG administration to pregnant women within 96 hours of exposure to chicken pox. 265 Importantly, even if the Enders Table is accepted without adjustment, as Professor Curtis did, it does no more than provide a basis for an extrapolation across a population of the effectiveness of VZIG. It cannot be taken, without more, to directly apply to Mrs King. At best, it is circumstantial evidence of the possibility that in the case of Mrs King, if VZIG was administered, she may have avoided an infection." (Red 292M - 293F) 94His Honour rejected the appellant's reliance on a "common sense approach". His Honour did not accept that guidelines which prescribed or recommended the use of VZIG provided proof of factual causation. He considered that such guidelines, or broad recommendations, across a population did no more than indicate steps taken for the protection of the population as a whole. He accepted that VZIG may well be effective across a population, but this did not establish causation with respect to the appellant. 95The primary judge referred to the following paragraphs from McGuiness [60], [78] - [80]. By reference to those statements of principle, his Honour concluded that the scientific evidence in the case before him established a possibility of prevention of infection but no more: "278 It is accordingly necessary to consider and determine whether there is any particular factor available on the whole of the evidence, which relates to Mrs King individually and which could form the basis of the conversion of a possibility which the evidence of Professor Curtis and the experts all support, to a probability that VZIG administered to her would have prevented her infection with VZV. 279 The plaintiff in her submissions did not refer to, or identify any fact, matter or circumstance in addition to those evidentiary facts and findings to which I have earlier made reference, which would support a conclusion that it was more probable than not that she would not have been infected with VZV if she had received VZIG." (Red 295W - 296F) Submissions and consideration 96The appellant submitted that the trial judge failed to give proper effect to the evidence as to the propensity of VZIG to prevent infection and that he failed to apply the correct legal test to the question of causation. 97In relation to the evidentiary issue, the appellant submitted that the Enders Table (adjusted or not) showed strong evidence of a cause and effect nexus in respect of a non-infection outcome when VZIG was administered to pregnant women. The appellant accepted that neither the Enders Table nor any other similar study, regardless of the percentage of infected and non-infected mothers, could say whether a particular plaintiff's mother would have avoided infection if administered VZIG. This would always be the case because of the general nature of such studies. 98The appellant submitted that the error implicit in his Honour's approach to the evidence was that he required the evidence to establish on the balance of probabilities an actual non-infection outcome for the mother. The appellant submitted that this imposed an incorrect evidentiary standard. By way of illustration, the appellant submitted that on the logic applied by his Honour, if the studies had shown a 75 percent non-infection outcome, she would still have failed because the mother could not on balance be placed into that 75 percent. 99The appellant submitted that the opinion of Professor Curtis (that on the balance of probabilities and informed by the Enders Table and other evidence, the administration of VZIG would have prevented the mother's varicella infection), should have been accepted and acted upon by his Honour. 100The appellant submitted that the fact that the Enders Table was an observational study, had only a small sample (212) and as such had a fairly low level of classification, was not significant. This was because it was the best study which was available and because of ethical considerations (i.e., the unacceptability of a refusal to administer VZIG in circumstances of a non-immune pregnancy) a higher level classification test could not be carried out. This was agreed by all the experts. They agreed that the Table was "the best available material from which to understand the effects of VZIG in pregnant women" (Red 275R). 101The appellant submitted that his Honour erred in taking into account the different dosage levels of VZIG received by the patients in the Enders Table and that which would have been administered to the mother in Australia, i.e., 1750 - 2000 international units in the Enders Table and 600 international units in Australia. The appellant submitted that the effects of an increased dosage were simply not known. No comparative test had been carried out. It was therefore speculative on the part of his Honour to take this matter into account. 102In the alternative, the appellant submitted that by reference to the date when she should have been administered VZIG, and the fact that VZIG should be administered as soon as possible for maximum effect, there was no evidence to suggest that the administration of 600 international units, had it been administered when it should have been, would have been any less effective than 1750 - 2000 as referred to in the Enders Table. Accordingly, the difference in dosage levels did not weaken the effect of the Enders Table figures. 103The appellant was critical of his Honour's use of the strike rate of varicella, i.e. that 10 to 15 percent of non-immune individuals would not be infected after significant exposure, such as household contact. The appellant submitted that his Honour should not have adjusted the Enders Table to take that into account. The appellant submitted that even if the strike rate were taken into account and the Table were adjusted, the results still meant that the number of patients with "no infection" was in the order of 50 percent. This was a significant number and did not invalidate Professor Curtis' reliance upon the Table. 104The appellant submitted that whether adjusted or not for the virus strike rate, the Enders Table demonstrated that a significant proportion of the pregnant women in the study population did not become infected if administered VZIG. Since causation and proof generally were not merely questions of numbers, there was little evidential significance between the Enders Table and the adjusted table. The appellant submitted that his Honour erred in attributing significance to the altered numbers on the adjusted table and their impact on the evidence of Professor Curtis. 105The appellant submitted that the Enders Table still demonstrated, as a study of limited scope but the best study available, that there was a substantial infection prevention outcome if VZIG were administered and even if the imprecise other adjustments were considered. That substantial infection prevention outcome should have properly resulted in a finding on the probabilities that the mother's infection would have been prevented, because it properly permitted a reasonable satisfaction by the tribunal of fact of the occurrence of the event in question: Briginshaw v Briginshaw [1938] HCA 34; 60 CLR 366 at 361 - 362 per Dixon J. 106In oral submissions (AT28.17 - 29.22), the appellant identified the following matters which were said to establish, on the balance of probabilities, that had on 6 May 2002 the medical staff at the Blacktown Hospital administered 600 international units of VZIG to the appellant's mother she would not have contracted varicella on 16 May 2002. (i) VZIG is a made-for-purpose product. (ii) There was no evidence that during the first 72-96 hours of the exposure to varicella, the virus particle numbers would have reached such a level that the VZIG antibodies would not have been able to deal with them. (iii) The failure to administer VZIG occurred well within the 96 hour prescribed period. (iv) The foregoing matters boded well for VZIG to have its designed outcome. (v) All of the experts would have recommended the administration of VZIG to the appellant's mother. (vi) Whatever the limitations of the scientific evidence, it did not demonstrate that VZIG was not efficacious to neutralise virus particles and result in a non-infection outcome. (vii) The appellant developed CVS and the mother was infected with varicella. 107The appellant submitted that those matters, together with the Enders Table, provided compelling circumstantial evidence and compelled a finding that on balance had VZIG been administered to the appellant's mother, a non-infection outcome would have been obtained. 108In order to assess the appellant's summation of the evidence, it is necessary to have regard to the purpose which the appellant sought to achieve by adducing expert evidence. As was stated in Strong v Woolworths Ltd [2012] HCA 5; 246 CLR 182: "18 The determination of factual causation under s 5D(1)(a) is a statutory statement of the "but for" test of causation: the plaintiff would not have suffered the particular harm but for the defendant's negligence... ... 32 The appellant was required to prove on the balance of probabilities that Woolworths' negligence was a necessary condition of her harm. Woolworths' negligence lay in its failure to employ a system for the periodic inspection and cleaning of the sidewalk sales area. Proof of the causal link between an omission and an occurrence requires consideration of the probable course of events had the omission not occurred. Here, the appellant was required to prove that, had a system of periodic inspection and cleaning of the sidewalk sales area been employed on the day of her fall, it is likely that the chip would have been detected and removed before she approached the entrance to Big W." 109Translating that test to the facts of this case the question becomes whether the omission to administer VZIG on 6 May 2002 was a necessary condition of the mother developing varicella. The answer to that question required a consideration of the probable course of events had the omission not occurred, i.e., had VZIG been administered on that date. What the appellant had to prove to establish causation was that had VZIG been administered on that day, it was likely that she would not have developed varicella. 110The way in which the appellant sought to establish that proposition at trial was through the evidence of Professor Curtis. Professor Curtis was the only expert to express the opinion in unqualified terms that on the balance of probabilities, the development of varicella in the mother would have been prevented had VZIG been administered on 6 May. 111None of the other experts was prepared to go that far. They all agreed it was possible. Initially Professor Turner was prepared to accept a 50 percent chance, but later qualified that evidence (Black 86P). Accordingly, it became necessary for the primary judge to assess the basis upon which Professor Curtis had formed that opinion. This required him to examine the reasons of Professor Curtis and the evidence of the other experts, where that evidence did not agree with or support that of Professor Curtis. 112It seems clear, as his Honour found, that the statistics contained in the Enders Table played an important part in the reasoning process which led to Professor Curtis forming the opinion which he did. At Blue 709V, which was a summary of responses to certain questions set out in the joint experts' report, Professor Curtis provided a note to justify his response as follows: "6th May 2002 was within 72 hours of exposure; table inserted above by BH for Q21 shows VZIG prevented infection in 54 percent of women given VZIG 1 - 2 - 3 days after exposure - therefore on the balance of probabilities the plaintiff's mother would not have." 113I have concluded that his Honour's reservations concerning the Enders Table were justified and that it does not constitute compelling circumstantial evidence, such as would justify the conclusion of Professor Curtis. 114The "study quality issue" should not be ignored. It is true that the Enders and Miller article, of which the Enders Table formed a part, is the best information we have. Nevertheless, it had a low "classification of evidence" level and involved a small cohort of 212 patients. Despite the fact that it is the best evidence which is available, the reliability of the results set out has to be qualified by those undoubted limitations. 115The "differential dosage" issue is also a matter which weakens the reliability of the results in the Table when applied to the circumstances in Australia in 2002. This is not a criticism based on "glimpses and snippets and a bit of conjecture and speculation", as was suggested in oral submissions (AT 54.50 - 55). The reservations were expressed by the authors and are to be found in the section of the article where the Table occurs. That part of the article states: "Management of varicella in pregnancy Passive prophylaxis with immunoglobulin Passive antibody prophylaxis of the standardised preparation of VZIG ... is recommended for exposed susceptible pregnant women to prevent or attenuate varicella. To the extent that VZIG completely blocks infection, it has the theoretical potential to provide the most benefit for mothers who are exposed during the first half of pregnancy, when the foetus is at risk of CVS. VZIG prophylaxis for women exposed later in pregnancy is recommended to reduce the risk of maternal pneumonia ... Two preparations of VZIG for intramuscular administration are produced in Scotland and the UK, and are distributed by the Public Health Laboratory Service. These preparations have a relatively low IgG antibody concentration of about 700 IU per adult dose. Their effect is to attenuate disease, rather than to prevent infection. VZIG preparations for intramuscular use are also available in the USA and Scandinavian countries. Subclinical infection occurs in a substantial proportion (15 percent) of high-risk nonimmune infants and children exposed to varicella and given UK VZIG as detected by testing for seroconversion in blood taken one-month post exposure. In Germany, two commercial preparations of VZV immunoglobulins defined antibody concentrations are available, one for intramuscular (IM) and another for intravenous (IV) administration, with an IgG antibody concentration of approximately 2000 IU per adult dose. According to our experience in 212 seronegative pregnancies, about half of the mothers have clinical varicella, even if VZIG is given in the recommended dosage IM or IV, within 1 - 3 days of significant exposure, with modified or normal disease, and a further 5 percent have subclinical infection (Table 16.7). ... Whether VZIG has any benefit in prevention of CVS is not known. Comparing experiences with VZIG administration during pregnancy is difficult because the quality (i.e. IgG antibody concentration) of the various VZIG preparations that are used worldwide (e.g. UK, Scandinavia, France, Germany, USA) is different. Furthermore, it should be realised that immunoglobulin preparations should be measured for neutralising antibody titres instead of determining the IgG concentration by ELISA. The nature of the maternal exposure is difficult to evaluate. No studies have been done with cohorts of pregnant women who had proven susceptibility to varicella, confirmed closed contact and included population sample sizes that are adequate to provide data about efficacy. ... This case demonstrates that a very large prospective body would be needed to clarify whether passive prophylaxis is beneficial for the prevention of CVS." (Emphasis added.) 116Evidence to similar effect was given by Professor Kesson and Dr Hudson as follows: "WITNESS KESSON: The effect of varicella zoster immunoglobulin at 600 international units per adult dose could be both to attenuate disease and to prevent infection. However, I think it is more likely to attenuate infection than prevent disease. ... WITNESS HUDSON: Can I just add to that? In my mind, this relates to the fact that the studies by Enders were conducted in Germany and used a dose of three times that amount of VZIG, three times that amount. And therefore it is known that the higher dose is more likely to prevent infection and therefore attenuation of disease is more likely to be seen with a lower dose than with a higher dose." (Black 108 G - K) To that should be added the evidence of Professor Kesson at [75] hereof. 117The effect of that evidence is that the antibody concentration in the dose which would have been administered to the appellant's mother in 2002 was considerably less than in the dose administered to the patients in the Enders Table. The effect of a lower antibody dose being administered is to attenuate, rather than prevent, varicella. That is important because prevention was that which was required for causation to be established in this case, not attenuation. 118If anything, his Honour's finding in this regard at Red 278I - L was conservative. The evidence on this issue is strongly suggestive of some doubt as to the applicability of the results as shown on the Enders Table to what would have happened in Australia in May 2002. His Honour was certainly entitled to conclude that at the very least, some of the patients recorded on the Table as showing no infection would in the Australian context, have been likely to be infected with varicella. Although as his Honour pointed out, it is not possible to know exactly to what extent the numbers of uninfected patients on the Table would have been reduced, the fact of a reduction in such a small cohort would have been statistically significant. 119That proposition was not challenged by Professor Curtis during the concurrent evidence. It is not answered by the appellant's submission that because the VZIG would have been administered well within the recommended 96 hours of exposure having taken place, it would therefore have been effective even at a dose of 600 international units. There was no expert evidence to that effect. On the contrary, the only evidence was to contrary effect. Accordingly, I agree with the conclusion of the primary judge at Red 278W that the Enders Table overstates the effectiveness of VZIG in varicella prevention when considering its application to the Australian administration practice. I would add, by reference to the qualification of the authors of the Table (see [115]) that the level of overstatement could be substantial. 120The final matter which the primary judge took into account was the "virus strike rate issue". It was common ground among the experts that in circumstances of significant exposure such as household contact, there was likely to be an 85 - 90 percent strike rate of infection with varicella. The corollary which was also accepted was that 10 - 15 percent of exposed non-immune patients would not be infected. It was also accepted by the experts that the Enders Table did not take this issue into account. 121The appellant challenged the use by his Honour of this evidence in two respects. The first challenge was that his Honour should not have adjusted the Table, but should have simply relied upon the evidence given. I do not understand that submission. The evidence was clear that the Table had not taken into account an important consideration. There was also evidence as to how that consideration could be taken into account. Accordingly, his Honour did not err in taking that consideration into account. 122The second challenge was that by adjusting the Table, his Honour had taken the consideration into account incorrectly. Professor Curtis agreed that it was appropriate to adjust the Table to take into account the virus strike rate issue but that the only adjustment which needed to be made was to the "non-infected" group. Having heard submissions on the issue, the primary judge concluded that Professor Curtis was incorrect in that approach and that, using a rate of 10 percent, the whole of the cohort of 212 patients should be adjusted to reflected the virus strike rate issue, i.e., that 21 of the cohort would not have been infected in any event, regardless of whether VZIG had been administered or not. 123I have concluded that the approach of the primary judge to the adjustment of the Table is to be preferred. On that approach, the non-infection rate of those who had received VZIG was less than 50 percent. Even on the approach of Professor Curtis, adjusted as he suggested, the non-infection rate was no greater than 50 percent. 124When one takes into account the small number of the cohort, the low level classification of the study, the significant difference in dosage levels between Australia and Germany, with particular reference to the necessity for prevention rather than attenuation of the disease, and the need to adjust the Table to take into account the "strike rate issue", the Table does not provide a persuasive basis for the opinion of Professor Curtis. The Table may be the best evidence available but its inherent deficiencies should not be ignored when assessing the opinion of Professor Curtis which substantially relied upon it. 125As the primary judge appreciated, however, the opinion of Professor Curtis was not based entirely upon the Enders Table. It was also based on other studies, which Professor Curtis acknowledged, examined the effects of VZIG when administered in quite different circumstances to that before the Court. No challenge was made in the appeal to his Honour's analysis and conclusions with respect to those other studies. His Honour's conclusions accord with my reading of the studies. 126The studies made no distinction between prevention and attenuation of varicella. Some related to newborn babies and some made no distinction in the cohort under consideration between patients who were adults and children. Some of the cohorts were very small, e.g. 12 in one case. Most importantly, however, there was no consistency in the results produced by the studies. As his Honour pointed out, those studies like the Enders Table, were also complicated by the potential for different dosages of VZIG in Australia and the countries where the studies were carried out. 127It follows that there is no basis for not accepting his Honour's conclusions in relation to those studies. In that regard, his Honour's general observation was: "I am unable to find in any of the expert evidence and other material dealing with these studies any explanation for the difference in the study results, save to note that it seems clear that VZIG works differentially in different groups of patients to whom it is administered." (Red 287 K - L) 128The final basis for the opinion of Professor Curtis was his clinical experience when dealing with severely immunocompromised children with cancer. Professor Curtis had found VZIG to be highly effective for such patients. Associate Professor Kesson, who had a similar practice, agreed with that conclusion. Her reservation was that the administration of VZIG to a pregnant woman to prevent, rather than attenuate varicella infection, was not an analogous situation. His Honour accepted that reservation. 129His Honour was entitled to do so. As his Honour noted (Red 291W - X) there was no general comparability between immunocompromised patients (such as those with whom professor Curtis was dealing) and pregnant women. On that issue, his Honour referred to the special circumstances of a pregnant woman, who is not "immunocompromised" in the conventional sense (see [91]). In those circumstances, the clinical experience of Professor Curtis did not provide a sufficient basis for concluding that VZIG was more likely than not to be effective in preventing varicella infection in pregnant women generally or in the appellant. 130For the appellant to establish causation the evidence of Professor Curtis, and in particular his reliance upon the Enders Table, had to be accepted. The evidence of the other experts, not only did not provide support, but was to the contrary. In addition, there was the recommendation of the Advisory Committee on Immunisation Practices of 2007 in the United States that: "Pregnant women. Because pregnant women might be at a higher risk for severe varicella and complications, VZIG should be strongly considered for pregnant women without evidence of immunity who have been exposed. Administration of VZIG to these women has not been found to prevent viremia, foetal infection, Congenital Varicella Syndrome, or neonatal varicella. Thus, the primary indication for VZIG in pregnant women is to prevent complications of varicella in the mother rather than to protect the foetus. Neonates born to mothers who have signs and symptoms of varicella from five days before to two days after delivery should receive VZIG, regardless of whether the mother received VZIG." (Blue 216 L - N) 131For the reasons indicated, the evidence of Professor Curtis did not establish the causal link between the hospital's omission to offer VZIG to the appellant and the development of her varicella. This was because the evidence did not establish that had VZIG been administered; it was likely that the appellant would not have developed varicella (Tabet v Gett [2010] HCA 12; 240 CLR 537 at [140]. 132The appellant challenged the primary judge's approach to causation on the basis that he had adopted a too mechanistic approach to the Enders Table. The appellant submitted that, having adjusted the Table in such a way as to indicate a less than 50 percent probability of non-infection with varicella had VZIG been administered, his Honour assessed causation on that basis. 133This submission does not do justice to the approach of the primary judge. 134The primary judge had two reservations concerning the Enders Table, both of which were justified. The first was its inherent weakness as an evidentiary basis for establishing the propositions set out therein, with particular reference to what was likely to occur in Australia. That has already been discussed. 135His Honour's second reservation was based on the Table's limitations as epidemiological evidence in that even if the data therein set out was accurate and relevant to Australia, it would establish a possibility at most, not a probability. His Honour's conclusions in that regard were well founded. 136In Merck Sharp and Dohme (Australia) Pty Ltd v Peterson [2011] FCAFC 128 the Federal Court (Keane CJ, Bennett and Gordon JJ) said on this issue: "106 It is desirable to begin by referring to the legal principles governing the scope of the legitimate use of epidemiological evidence in relation to proof of causation. We begin with the proposition that proof of what may be expected to happen in the usual case is of no value unless it is proved that the particular applicant is indeed "the usual case". In State Government Insurance Commission (South Australia) v Laube (1984) 37 SASR 31 at 33 King CJ said: "The use of statistical probabilities in legal proof has been the subject of much learned writing... In the article referred to above, in [1979] Crim. L.R. at p 305 Professor Glanville Williams refers to a rule of law relating to proof "that evidence should focus on the defendant". ... I am clearly of the opinion that the statistical fact that a particular proposition is true of the majority of persons cannot of itself amount to legal proof on the balance of probabilities that the proposition is true of any given individual. 107 In Seltsam v McGuiness, Spigelman CJ examined at length the significance of epidemiological evidence for the resolution of the issue of causation. He concluded (at [89] and [98]) that epidemiological studies are evidence of possibility which may, alone or in combination with other evidence, establish causation, in a specific case, on the balance of probabilities. Spigelman CJ said at [118]-[120]: "The issue in the present case is whether an increased risk did cause or materially contribute to the injury actually suffered. There is a tension between the suggestion that any increased risk is sufficient to constitute a "material contribution", and the clear line of authority that a mere possibility is not sufficient to establish causation for legal purposes. The latter is too well established to be qualified by the former. The reconciliation between the two kinds of references is to be found in the fact that, as in Chappel v Hart and in the cases that suggest the former, the actual risk had materialised. The "possibility" or "risk" that X might cause Y had in fact eventuated, not in the sense that X happened and Y had also happened, but that it was undisputed that Y had happened because of X. The epidemiological evidence in the present case can be expressed in terms of "increased risk". However, in its application to determining causation in the specific case of the respondent that evidence never rises above the level of a possibility. Whether or not the increased risk "eventuated", is the issue which must be determined. The respondent's reliance on the passage from McHugh J was, in my opinion, misplaced. ... 109 Having reviewed the case law in the United States, Spigelman CJ concluded at [135]- [137]: "Some of the American cases indicate that the [relative risk] of 2.0 should not be applied as a rigid mathematical formula. Others appear to apply it in that way. The predominant position in Australian case law is that a balance of probabilities test requires a court to reach a level of actual persuasion. This process does not involve a mechanical application of probabilities ... In Australian law, the test of actual persuasion does not require epidemiological studies to reach the level of a relative risk of 2.0, even where that is the only evidence available to a court. Nevertheless, the closer the ratio approaches 2.0, the greater the significance that can be attached to the studies for the purposes of drawing an inference of causation in an individual case. The "strands in the cable" must be capable of bearing the weight of the ultimate inference." 110 A difficulty, as a matter of legal policy in the way of adopting the rule of thumb adopted in the United States cases, is that it is apt to mandate an award of compensation to applicants who have not, in truth, been injured by the respondent. That is because those applicants who were actually injured by causes other than the respondent's actionable conduct will be able to recover compensation because, for them too, a relative risk of greater than 2 can be said to imply probability of greater than 50% that the respondent's actionable conduct was the cause of their loss. 111 It should also be borne in mind that, while a relative risk of 2 might imply a 50% probability that the risk has come home in a typical case, a relative risk of less than 2 would imply a probability of less than 50%, that is to say less probable than not. The primary judge's finding of relative risk was of "about 2". And in any event the strength of that finding as a strand in Mr Peterson's case is problematic because of the other candidates as causes of his injury. Further, the absence of a clear appreciation of the level of absolute risk to Mr Peterson from the actionable conduct of Merck and MSDA in comparison with those other candidates detracts from the force of the circumstantial case which he seeks to mount." 137This issue was considered by the High Court in Amaca Pty Ltd v Ellis; The State of South Australia v Ellis; Millennium Inorganic Chemicals Ltd v Ellis [2010] HCA 5; 240 CLR 111. There the Court said: "63 As explained at the commencement of these reasons, there being no direct evidence about what actually caused Mr Cotton's cancer, it was the plaintiff's case that the epidemiological evidence established facts which "positively suggested, that is to say provided a reason ... for thinking it likely" that, in exposing Mr Cotton to respirable asbestos fibres, the negligence of each defendant was a cause of his cancer. To draw an inference about causation from what was established by the epidemiological studies, it would be necessary to decide whether the particular case under consideration should be treated as conforming to the pattern described by the epidemiological studies. Absent evidence which suggests that the individual may stand apart from the ordinary, there may be sufficient reason to assume conformity, but whether or not that is so, it is important to recognise that the first step that must be taken, if an inference is to be drawn from epidemiological studies, is to relate the results of studies of populations to the particular case at hand. That step is not inevitable. 63 In this case, even if it were decided that Mr Cotton's case was not atypical, application of either the relative risk or the probability analyses given in evidence does not found the inference that the plaintiff sought to have made. The probabilities and distributions recorded in the epidemiological studies would point away from the conclusion urged by the plaintiff." 138Far from adopting a mechanistic or probabilistic approach, his Honour was mindful of the limitations of the Enders Table and in particular, its applicability to Australia. As the difference in dosage between Australia and Germany illustrated, it was by no means clear that the appellant constituted a "typical" member of the cohort on which the Table was based. That cohort comprised persons to whom a dosage of between 1750 and 2000 international units had been given, whereas the appellant was a person to whom a dose of 600 international units should have been given. Even looked at on a probabilistic basis, the Table established at best a 50 percent probability but more likely something significantly below that level. 139By referring to the general nature of the information provided by the Enders Table, his Honour was doing no more than stating as a matter of principle that the appellant had to show how that general statistical information applied to her particular case. He was not applying a mechanistic or probabilistic approach. 140The difficulty which his Honour identified for the appellant in moving from the general information provided by the Enders Table to establishing the likelihood of VZIG preventing the development of varicella in her case is not overcome by the seven factors identified in submissions (AT 28.21 - 29.15). The evidence as to when and to what extent the appellant was exposed to varicella, and therefore where along the 96 hour spectrum she would have been when administered VZIG, is simply unknown and too imprecise to draw any conclusion other than that had VZIG been administered on 6 May it would have been within the recommended 96 hour window of exposure for it to be efficacious. 141Similarly, the fact that all of the experts would have recommended the administration of VZIG to the appellant does not advance the issue of causation. It does not follow that the experts agreed that if administered VZIG was likely to have been effective in preventing the contraction of varicella. What it indicates is recognition by the experts that the contraction of varicella by a pregnant woman can have serious consequences for both herself and her child so that where there was a chance that that risk could be avoided by the administration of VZIG, that chance should be taken. Given the potentially serious consequences of the development of varicella in those circumstances, it would be unethical not to administer VZIG when there was a chance of those consequences being avoided. This does not establish that those consequences were likely to have been avoided. As the primary judge appreciated, this was a public health issue, not one related to causation (Red 293R 294N). 142The alternative basis upon which the appellant challenged the decision of the primary judge was that he did not apply the correct legal test to the question of causation. The appellant relied upon the statement of principle by McHugh J in Chappel v Hart [1998] HCA 55; 195 CLR 232 at [27] - [28] where his Honour said: "27 Before the defendant will be held responsible for the plaintiff's injury, the plaintiff must prove that the defendant's conduct materially contributed to the plaintiff suffering that injury. In the absence of a statute or undertaking to the contrary, therefore, it would seem logical to hold a person causally liable for wrongful act or omission only when it increases the risk of injury to another person. If a wrongful act or omission results in an increased risk of injury to the plaintiff and that risk eventuates, the defendant's conduct has materially contributed to the injury that the plaintiff suffers whether or not other factors also contributed to that injury occurring. If, however, the defendant's conduct does not increase the risk of injury to the plaintiff, the defendant cannot be said to have materially contributed to the injury suffered by the plaintiff. That being so, whether the claim is in contract or tort, the fact that the risk eventuated at a particular time or place by reason of the conduct of the defendant does not itself materially contribute to the plaintiff's injury unless the fact of that particular time or place increased the risk of the injury occurring. 28 In principle, therefore, if the act or omission of the defendant has done no more than expose the plaintiff to a class of risk to which the plaintiff would have been exposed irrespective of the defendant's act or omission, the law of torts should not require the defendant to pay damages. Similarly, if the defendant has done no more than expose the plaintiff to a risk for which the defendant has not undertaken responsibility and to which the plaintiff was always exposed, the law of contract should not require the defendant to pay damages for injury arising from that risk even if it follows upon a breach of contract. No principle of the law of contract or tort or of risk allocation requires the defendant to be liable for those risks of an activity or course of conduct that cannot be avoided or reduced by the exercise of reasonable care unless statute, contract or a duty otherwise imposed by law has made the defendant responsible for those risks." 143The appellant submitted that the hospital's conduct in not offering to administer VZIG on 6 May increased the risk of injury to her. The risk which was increased was that she would develop varicella. She did in fact develop varicella and accordingly, the risk of injury came home in the relevant sense and causation was established in that the hospital's conduct materially contributed to the injury. 144On that latter issue, the appellant called in aid the analysis of material contribution in Strong v Woolworths Ltd at [23] - [29]. 145The appellant did not explain how the analysis of material contribution in Strong v Woolworths Ltd assisted her case. As I read the analysis of the plurality, there are real difficulties in applying the "but for" test to the concept of "increase in risk". The reference to and analysis at [24] - [27] of Bonnington Castings Ltd v Wardlaw (1956) UKHL 1; (1956) AC 613 illustrates the difficulty. Far from assisting the appellant's submission, the analysis in Strong v Woolworths of "increase in risk" in the context of a "material contribution to harm" is strongly suggestive that in most of those cases, the "but for" test will not be satisfied and that recourse will have to be had to s5D(2) CLA. 146There is another difficulty with the appellant's articulation of the basis for establishing causation relying on Chappel v Hart. The negligence on the part of the hospital was one of omission in failing to offer a particular therapeutic substance. As Strong v Woolworths Ltd established "Proof of the causal link between an omission and an occurrence requires consideration of the probable course of events had the omission not occurred". 147Here the risk to which the appellant was exposed existed regardless of any conduct of the hospital, i.e. she had already been exposed to the varicella virus. It is therefore not correct to say that the hospital had "increased the risk of injury" and that this risk had eventuated. To apply that test to the facts of this case involves a circuity of reasoning. It presupposes an affirmative answer to the fundamental causation question. 148In the appellant's submission the risk created by the hospital's omission presupposes that the appellant was one of those persons in relation to whom the administration of VZIG was likely to prevent varicella. It, in effect, assumes an affirmative answer to the very question posed by the evidence in the factual causation enquiry previously discussed, i.e. the likelihood of VZIG if administered to the appellant on 6 May being effective to prevent varicella infection. 149To apply the Chappel v Hart test of causation, in the way in which the appellant seeks to do, would always result in an affirmative result for a plaintiff in a case such as this regardless of whether there was but a very low prospect of infection being avoided. The infected person would always be able to say that the risk of infection had been increased by the failure to administer the therapeutic substance and therefore the risk had come home. 150This was the very issue identified by Spigelman CJ in Seltsam Pty Limited v McGuiness where his Honour said: "105 The respondent relied on an observation by McHugh J in Chappel v Hart supra [272] where his Honour, noting that "increases" in this context includes "creates", said: "If a wrongful act or omission results in an increased risk of injury to the plaintiff and that risk eventuates, the defendant's conduct has materially contributed to the injury that the plaintiff suffers whether or not other factors also contribute to that injury occurring. If, however, the defendant's conduct does not increase the risk of injury to the plaintiff, the defendant cannot be said to have materially contributed to the injury suffered by the plaintiff." 106 Although his Honour's was a dissenting judgment, this passage has subsequently been referred to with approval (see Naxakis v Western General Hospital [1999] HCA 22; 73 ALJR 782 at [31] per Gaudron J and [127] per Callinan J. 107 The starting point of McHugh J's analysis was that it had been established on the balance of probabilities that the conduct did create or increase the risk of injury, "and that risk had eventuated". 108 This starting point is the very matter in issue in the present case. Was there evidence on the basis of which the trial judge could conclude, on the balance of probabilities, that there was an increased risk of injury and that that risk had "eventuated" in the specific disease of the respondent. ... 118 The issue in the present case is whether an increased risk did cause or materially contribute to the injury actually suffered. 119 There is a tension between the suggestion that any increased risk is sufficient to constitute a "material contribution", and the clear line of authority that a mere possibility is not sufficient to establish causation for legal purposes. The latter is too well established to be qualified by the former. The reconciliation between the two kinds of references is to be found in the fact that, as in Chappel v Hart and in the cases that suggest the former, the actual risk had materialised. The "possibility" or "risk" that X might cause Y had in fact eventuated, not in the sense that X happened and Y had also happened, but that it was undisputed that Y had happened because of X." 151The correct approach was stated succinctly by Mason P in TC by his tutor Sabatino v State of New South Wales & Ors [2001] NSWCA 380 where his Honour said: "59 I cannot accept this submission. I remain of the view that Australian law has not adopted a formal reversal of onus of proof of causation in negligence, even negligence involving breach by omission. A robust and pragmatic approach to proof of causation permits, but does not compel, a finding of liability in cases of negligence by omission which (as Gaudron J points out in Bennett) is necessarily based upon a hypothetical enquiry. A defendant who exposes a plaintiff to a risk of injury or who, by omission, fails to take reasonable steps to avoid or minimise that risk is not liable unless the risk comes home in the sense that the court is ultimately satisfied on the balance of probability that the defendant's breach caused or materially contributed to the harm actually suffered." See also Flounders v Millar [2007] NSWCA 238 at [1] - [40] (Ipp JA with whom Handley JA and Hoeben J agreed). 152The alternative basis of liability submitted by the appellant based on Chappel v Hart does no more than drive the inquiry back to the first point raised by the appellant in the appeal, i.e. the factual inquiry as to whether the appellant proved that had she been administered VZIG on 6 May, it was likely that she would not have been infected with varicella. 153Despite senior counsel for the appellant expressly eschewing any reliance upon s5D(2) CLA at trial, the appellant sought to rely upon the section in the appeal. This occurred in oral submissions at AT35.28. The submission was that if it be said that the Chappel v Hart approach does not fall within s5D(1), this would give rise to an "exceptional case" such as would come within s5D(2). That submission was not further developed. In that regard, the appellant did no more than to make a general reference to the discussion by the plurality in Strong v Woolworths Ltd at [22] - [29]. 154The appellant's grounds of appeal do not include an application to rely upon s5D(2), nor was leave formally sought to advance a case on appeal which was not advanced at trial. In the circumstances of this case, the appellant should not be allowed to raise s5D(2) in passing, and then expect this Court to raise for itself the competing submissions and reach a conclusion. The Court is entitled to expect more assistance than this. 155As the discussion in Strong v Woolworths Ltd makes clear, the application of s5D(2) involves an examination of the Ipp Report, an analysis of the causation issues discussed by Professor Stapleton and the application of normative considerations to circumstances which may or may not constitute "an exceptional case". The normative considerations may require the adducing of evidence. Where it is intended to ask this Court to embark upon such an inquiry, the matter needs to have been raised at trial and the Court is entitled to the benefit of full argument. It follows that this Court should not undertake an inquiry as to the application of s5D(2) CLA in these circumstances. 156The appellant's submissions on causation were put at a level of generality and did not relate specifically to any of the grounds of appeal. For the reasons set out above, I have concluded that the appellant's challenge to the primary judge's findings as to causation fail and that the appeal should be dismissed. Nevertheless, it is still necessary to deal with the particular grounds of appeal lest an issue which should have been considered by the Court has not been taken into account. 157Grounds of Appeal 1 - 4 raise general issues which have been addressed above. 158Ground 5: His Honour erred in considering the opinion of Professor Curtis on a question of causation, in failing to take into account the opinions of Associate Professor Kesson (at J187 and 192) and the evidence of Dr Hudson (at J189) The point sought to be made by this Ground of Appeal is not altogether clear. The very fact that the primary judge set out the opinions of Associate Professor Kesson and Dr Hudson in the paragraphs nominated makes it clear that he did take them into account and did so as part of his assessment of the opinion of Professor Curtis. This Ground of Appeal has not been made out. 159Grounds of Appeal 6, 7 and 8 have been addressed above. 160Ground 9: His Honour erred in that he failed to add the percentage of persons for whom damage would have been ameliorated to those for whom infection would have been prevented before determining whether or not on the balance of probabilities administration of VZIG would have been effective. Ground 10: His Honour erred in that having found that on the balance of probabilities the administration of VZIG would not have prevented chickenpox in the plaintiff then failed to consider whether it would have ameliorated the effects of chickenpox. No submissions were directed to these Grounds of Appeal. There was no evidence at trial to the effect that amelioration, rather than prevention, of infection with varicella would have made any difference to the development of CVS in the appellant. The trial was run on the basis that in order to establish causation, the appellant needed to establish on balance that the administration of VZIG on 6 May to her mother was more likely than not to have prevented her becoming infected with varicella. These grounds of appeal have not been made out. Conclusion 161The orders which I propose are: (1) Leave to appeal is granted. (2) The appeal is dismissed. (3) The appellant is to pay the respondent's costs of the appeal. 162WARD JA: I have had the advantage of reading in draft the judgments of each of Basten and Hoeben JJA. For the reasons set out below, I agree with Hoeben JA that the appeal should be dismissed with costs. 163The facts giving rise to this appeal are set out in the judgment of Hoeben JA. The relevant question on this appeal is one of causation. There is no dispute as to the existence of a duty of care owed to the appellant to advise her mother (Mrs King) in relation to the availability of VZIG (varicella-zosta immuno-globulin), a biologic drug, for treatment following Mrs King's exposure, while pregnant with the appellant, to the varicella (chickenpox) virus. Nor is there any dispute that the respondent breached that duty of care in failing so to advise Mrs King, with the result that VZIG was not prescribed within the time frame within which VZIG would (in accordance with standard medical practice in Australia) have been advised to be administered. (I note that there was some reference at Blue 235 to there being a shortage of the drug but have assumed, there being no suggestion by the respondent to the contrary and the matter not having been raised in the proceedings, that had the drug been prescribed it would have been available to be administered within the 96 hour 'window'.) 164The appellant developed FVS (foetal varicella syndrome), also known as congenital varicella syndrome (CVS), as a result of her mother's varicella infection and is disabled as a consequence. 165Expert evidence was relied upon to establish whether, had VZIG been administered to Mrs King when she attended the hospital after her exposure to the virus, Mrs King would not have developed varicella. Had Mrs King not developed varicella (or a sub-clinical infection) from her exposure to the virus, there is no doubt that the appellant would not have been exposed to the risk of contracting FVS. It seems to have been accepted (AT 17.35; 31.8; judgment at [75] and [249]) that Mrs King having contracted varicella, there was a 1-2% risk that the virus would be transmitted to the foetus (although there was some reference in the evidence that the incidence of FVS in Europe where mothers were exposed to varicella in the first trimester was 0.4%, which suggests that the risk may have been even lower than that on which the argument below proceeded). Mrs King's evidence was that when she was diagnosed with chicken pox she was advised that there was less than 1% risk of the foetus being affected (Blue 421). 166The primary judge has been criticised by the appellant for adopting an overly mechanistic analysis of the scientific evidence, a criticism which Basten JA considers has some force. I am not persuaded that the primary judge's consideration of the statistical evidence put before him was overly mechanistic. His Honour considered not only the statistical evidence (the limitations of which were expressly noted) but also the evidence of the medical practitioners by reference to their clinical experience. Having taken those matters into account, the primary judge was not satisfied on the balance of probabilities that causation had been established. 167It was submitted by Mr Romaniuk, Counsel appearing for the appellant, that there was a need for his Honour to "go beyond" the epidemiological studies and to ask himself whether there was actual persuasion. I see no basis for concluding that his Honour did not do so. While there was criticism made of his Honour's approach (in posing the question as to within which of the groups in the study reflected in the Enders Table, to which I refer below, Mrs King was likely to have fallen), I consider the conclusion his Honour reached as to causation was correct. 168I agree with Basten JA (at [23]) that the ultimate question as to causation in the present case is not whether or not administration of VZIG to the mother would have prevented her developing adult chickenpox after exposure to the varicella virus (or would have attenuated the effects of the infection in the mother) but whether or not the failure to offer or recommend (and administer) VZIG to the mother materially contributed to or caused the occurrence of FVS in the foetus. (As to the meaning of material contribution to the harm, in Zanner v Zanner (2010) 79 NSWLR 702 at [11] Allsop P, as his Honour then was, noted that the concept of cause at common law can incorporate 'materially contributed to' in a way which would satisfy the 'but for' test and that "[s]ome factors which are only contributing factors can give a positive 'but for' answer".) 169The relevant assessment to be made in the present case was as to the probability of the appellant being affected by FVS, as a result of her mother's exposure to the virus, had VZIG been administered to the mother at the relevant time (and not the probability of the mother being affected by chickenpox). 170In Roads and Traffic Authority v Royal (2008) 82 ALJR 870 Gummow, Hayne and Heydon JJ said: In short, even if it could be said that the appellant's breach of duty "did materially contribute" to the occurrence of an accident, "by creating a heightened risk of such an accident" due to the obscuring effect of one vehicle on another in an adjoining lane, it made no contribution to the occurrence of this accident." (emphasis in original) 171Adapting the present circumstances to the language used in Royal, the question is not whether the failure to administer VZIG contributed to the development of adult chickenpox in the mother ("an" accident in the context of the above extract) but whether it made a material contribution to the contraction of FVS by the appellant ("this" accident). 172The emphasis put by the appellant is on the material non-infection outcome (whether adjustment be made to the Enders table or not) associated with the administration of VZIG and the "gut feeling" of the experts that if there was a chance of preventing a serious syndrome (FVS) that chance should be taken. Mr Romaniuk placed weight on the lack of clinical evidence that VZIG did not "perform". 173In answering the relevant causation question (and at the risk of attracting the same criticism as the primary judge), what must be considered is the material on which the parties relied (the appellant for the contention that she had suffered damage as a result of the hospital's breach of duty; the respondent for the contention that the causal link had not been established). That material comprised both the scientific evidence and the clinical experience of the relevant witnesses. 174In considering the scientific studies, I note that it was accepted that the quality of the studies put before the trial judge was quite poor by scientific standards (level III) but that this was the best available evidence. 175Of the scientific studies, the Enders study (which the primary judge considered in some detail) seemed to be accepted as the best available when assessing the likelihood of varicella developing after exposure to the virus if VZIG was administered within the relevant time frame. There was some doubt as to whether the studies reported in the table were confined to cases where VZIG had been administered in the German dosage (of 1750 International Units) or had included cases where a lesser dose had been administered. As a German study it seems likely that the cases sampled were those in respect of which the German dosage would have applied. 176Had VZIG been administered to Mrs King, the standard Australian dose that would have been administered would have been 600 International Units of VZIG. The appellant's case at trial was that administration of VZIG to Mrs King (at the standard Australian dosage) would have prevented the appellant developing FVS. Relevantly, the appellant's case at trial was not that administration of the standard Australian dose of VZIG would have attenuated the effects of the virus in the mother and thereby minimised or prevented the risk of the virus passing to the appellant. Rather, the case put by the appellant was that administration of the drug would have prevented the mother developing chickenpox following exposure to the virus, which would mean that the foetus could not then contract FVS as a consequence of the mother's chickenpox. I accept that there was discussion in the course of the concurrent evidence as to the effects or potential effects of VZIG to attenuate the effects of the disease in the mother. However, that was not the appellant's claim (which, proceeded on the basis that VZIG if administered had a 54% chance of preventing the mother developing varicella and thereby removing any risk that the appellant would develop FVS). (There was also some discussion by the experts as to the pathogenesis of the virus and, in particular, its ability to reactivate at a later time during the pregnancy when the placenta did not place a barrier on transmission of the virus to the foetus. However, it is not necessary to consider the implications of this for the present case.) 177At the outset, I note that the authors of the Enders study (Enders and Miller), writing in their textbook as to the efficacy of VZIG (Blue 265) state that "[w]hether VZIG has any benefit in prevention of CVS is not known" (at [337]), in part due to different preparations of the drug between countries and different dosages between different countries (see also Black 12.30) and in part due to the lack of high quality studies on this issue. 178Mr Romaniuk submitted that, on the results shown in the Enders study, had the appellant's mother been given the equivalent of the German dose of the drug (1750 International Units) she would have had a 54% chance of disease prevention. However, had the drug been administered to the appellant's mother in Australia, it would not have been at the German dosage. The "dosage issue", so-called, related to the lack of information as to the efficacy of differential dosage: in lay terms, it was not clear whether administration of a dose of VZIG nearly three times as much as the approved Australian dosage would have been more effective to prevent the mother contracting adult chickenpox or whether the Australian dose would have been sufficient to do so. There was also an issue as to the calibration of the drug, to which I refer below. 179Experts called to give evidence by the respondent, Professor Kesson (at Black 108G) and Dr Hudson (at Black 108K), agreed that dosage of VZIG at the Australian standard level would not have been preventative of the adverse effects of the disease (or otherwise curative of the disease, without negative adverse effects), but would merely have attenuated the effects of the disease in the mother. 180The expert called to give evidence by the appellant (Professor Curtis) addressed the distinction between prevention and attenuation of the virus (at Black 153T-154R). Professor Curtis confirmed that there are no studies which conclude at what point a dose becomes more likely to be attenuative (as opposed to preventative), noting that even if such studies did exist, the drug is materially different between countries. 181Professor Curtis also observed that no witness at the trial was an expert in VZIG 'calibration', as it was so called. His opinion was that the difference between the standard dose administered in other countries and that administered in Australia was not one that could be resolved by the expert evidence before the trial judge. 182The word 'calibration' may not adequately convey the difficulty of dosage comparisons. As I understand it, VZIG is human immuno-globulin containing high levels of anti-varicella antibodies. It is not a synthetic compound (such as paracetamol), where there should be chemical identity between the generic active ingredient of the drug produced by different manufacturers in different countries. As VZIG is biological in nature, one might expect that comparisons of the drug produced under different manufacturing arrangements would be more difficult. Nothing, however, turns on this in the present case, since there was no evidence as to its import on the assessment of the probability of the foetus contracting FVS. 183Another of the issues before the primary judge was what account, if any, was to be taken for the 'strike rate' of the disease, by reference to the proportion of those people who would not be adversely affected following exposure to the virus irrespective of whether VZIG had been administered. The strike rate was said to be 85-90% (i.e., of 100 people exposed to the virus, some 10-15 would not develop varicella whether or not VZIG was administered). This means that of the 54 people who, on the Enders study, will not develop the disease, statistically it will only be in the case of 39-44 of them that the non-infection is attributable to the administration of VZIG. Taking into account the strike rate of the disease (by eliminating from the statistical analysis the number that would not develop varicella whether or not the drug had been given), the primary judge considered that the administration of VZIG would reduce the risk of infection on the German (Enders) data to between 44% and 39%. If, as Basten JA explains, the number of people who would not in any event contract the disease needs to be excluded altogether from the analysis, then the risk of infection following administration of VZIG (at least in the dosage and at the calibration applicable in the Enders study) would be between 48% and 45% (depending on whether the strike rate was 90% or 85%). 184The need to account for the 'strike rate' arises because attributing to the success of VZIG those who would not have been adversely affected regardless of whether VZIG had been administered would incorrectly inflate the statistical success of the VZIG drug and the estimated probability that it would have prevented development of the disease in the plaintiff's mother. 185The Enders table was relied upon as relevant to the assessment of the probability of the mother developing varicella, following exposure to the virus, after the administration of VZIG (arguably, at the German dosage and calibration if the study was limited to German cases). As already noted, there was recognised to be a 1-2% chance of the appellant suffering any serious consequences as a result of her mother's chickenpox without administration of VZIG to the mother within the requisite 96 hour timeframe from the mother's exposure to the virus (as recognised by his Honour at [75]). 186To establish factual causation for the purpose of the appellant's claim (namely that the failure of the hospital to recommend the administration of VZIG to the mother was a necessary condition of the appellant developing FVS), it was necessary for the appellant to establish, on the balance of probabilities, that: her mother had contracted chickenpox within 96 hours of visiting the hospital (as was found by the primary judge and is not now disputed); her mother would have chosen to have VZIG administered had it been offered to her, as it should have been (again, as was found by the primary judge); and that administration of the Australian standard dosage of the drug as it was available in Australia at the relevant time to her mother would have prevented the appellant from contracting FVS as a result of her mother's infection with the varicella virus. 187The last of those steps (that his Honour was not satisfied had been established on the balance of probabilities), taking into account the statistics from the Enders study (and assuming for present purposes that there was no difference in the efficacy of the drug if administered at the Australian dosage calibrated in the Australian manner, as opposed to the higher dosage of the drug administered in Germany calibrated in the German manner, where it seems likely that the study was conducted), can be illustrated by the following statistics. 188If, using the figures as adjusted by his Honour, VZIG is statistically effective to prevent development of varicella in 39-44% of the relevant cohort, and only 1-2% of foetuses will contract FVS even if the mother is infected, then the administration of VZIG to the mother would have decreased the appellant's chance of being affected by FVS from 2% (taken at its highest) to 1.12-1.22% (1-0.39=0.61 0.02x0.61; 1-0.44=0.56 0.02x0.56) or (taken at the 1% figure, the lower of the range found by the primary judge) from 1% to 0.56-0.61%. (If, indeed, the risk was as low as 0.4%, as the later evidence suggested it might be, then the administration of VZIG would have decreased the risk to the foetus of developing FVS from 0.4% to 0.244-0.224%.) 189If the relevant percentage risk to the mother is 45%-48% (not 39% to 44%) on the adjusted Enders Table, then the risk to the foetus taken at the 2% figure would be reduced to 1.04-1.1% (1-0.45=0.55 0.02x0.55; 1-0.48 0.02x0.52) and taken at the 1% figure the risk to the foetus would be reduced to 0.52-0.55% (0.01x0.55; 0.01x0.52). 190In other words, the appellant's contention that administration of VZIG to the mother would more likely than not have prevented the appellant developing FVS must be seen against the statistic that VZIG (had it been administered to and effective on her mother) would have made the appellant under 1% less likely to be adversely affected by her mother's exposure to the chickenpox virus than she was without administration of the drug to her mother. 191The relevant "risk of harm", as noted earlier, was that to the foetus (not the mother). On the Enders study (adjusting for the "strike rate" of the disease), of 10,000 mothers exposed to the chickenpox virus, 8,500 to 9,000 will have adverse symptoms; of those, 3,900 to 4,400 (or, on the alternative adjusted figures, 4,600 to 4,800) will be prevented from suffering the adverse effects of chickenpox by the administration of VZIG (at least where and when the dose administered is consistent with that used in the study, as to which there was some uncertainty). 192However, the import of the empirical studies (to which Basten JA has referred, starting at [23]) relates to the "strike rate" of FVS in foetuses whose mothers have developed chickenpox. That may be explained as follows (looking at a different cohort from that to which I have referred in the paragraph above is that, out of 10,000 foetuses whose mothers develop chickenpox, between 100 (1%) and 200 (2%) will suffer FVS (at varying levels of seriousness) as a result, without the administration of VZIG. Of 10,000 foetuses whose mothers have been exposed to the varicella virus and whose mothers have been administered VZIG within the 96 hour time frame (at a sufficient dosage/calibration to prevent development of chicken pox in the mother), somewhere between 112-122 (on the 2% end of the statistical range) or 56-61 (on the 1% figure) will nevertheless suffer adverse affects (at varying levels of seriousness) related to FVS. 193If the effect of administration of VZIG to the mother would be sufficient only to attenuate the effects of the virus in the mother (rather than to prevent development of chickenpox in the mother), then it is not clear on the evidence what reduction, if any, there would be on the prospect of the foetus suffering FVS or the level of injury that would ensue if the foetus did. 194It is important to adumbrate the statistics into meaningful numbers (as above) to understand the import of the other studies identified by Professor Curtis, all of which had relatively small sample sizes. The study in The Lancet (Vol 343 pp1548-1551, 18 June 1994), for example, stated that "no cases of congenital varicella syndrome or zoster in infancy occurred among the 97 pregnancies in which maternal varicella infection followed post-exposure anti-varicella-zoster immune-globulin prophylaxis" (at page 1550). However, on the 2% figure referred to above, only about 1.08-1.18 children in the study would be expected to present with FVS symptoms (or on the 1% figure, only 0.54-0.59 children would be expected to present with FVS symptoms). The fact that no children did present with FVS is accordingly of little statistical significance. 195If the appellant's chances of being negatively affected by FVS could possibly have been ameliorated by less than one-percent (on either the 2% figure or 1% figure) by the administration to her mother of VZIG at the Australian dosage/calibration (whether or not that was sufficient only to attenuate the effect of the disease), then it is difficult to accept that the failure to prescribe the administration of VZIG caused or materially contributed to the appellant's injury and her claim must fail. 196This does not mean that a doctor, who knew that there was a potential decrease in the chance of harm to the foetus of less than 1% (with negligible side effects to the mother and foetus) should not have recommended administration of the drug, assuming it was available within the relevant time frame. The public interest in removing or reducing even an already small risk would obviously need to be taken into account. It simply means that factual causation was not established. 197As to the clinical experience of the medical practitioners who gave evidence, only Professor Curtis considered that it was more likely than not that the administration of VZIG would have been effective to prevent development of varicella in the mother. Professor Curtis' clinical experience was that VZIG was highly effective in preventing chickenpox in severely immunocompromised children. Whether that was able to be translated to the position of pregnant women (and, more relevantly, their foetus/es was not clear). Professor Kesson, for example, considered that pregnant women were not necessarily immunocompromised with respect to developing varicella. Professor Curtis agreed it was a difficult issue. 198The fact that the relevant risk was of FVS to the foetus and that, by not giving VZIG, the very kind of thing that the relevant duty obliged the hospital to take reasonable steps to prevent in fact happened (i.e. the foetus being afflicted by FVS), does not of itself answer the test for factual causation. The failure to administer VZIG must have caused or materially contributed to the injury. In Adeels Palace Pty Ltd v Moubarak (2009) 239 CLR 420 (and see also the more recent judgement of Tabet v Gett (2010) 240 CLR 537 at [66] per Hayne and Bell JJ, [114] per Kiefel J), the High Court said at [50]-[53]: Recognising that changing any of the circumstances in which the shootings occurred might have made a difference does not prove factual causation. Providing security at the entrance of the restaurant might have delayed the gunman's entry; it might have meant that, if Mr Bou Najem was a random victim, as seemed to be the case, someone else might have been shot and not him. But neither plaintiff proved factual causation by pointing to possibilities that might have eventuated if circumstances had been different. Nor was "but for" causation established in these cases by observing that the relevant duty was to take reasonable care to prevent injury to patrons from the violent, quarrelsome or disorderly conduct of other persons. That is, the question of factual causation was not answered in these cases by pointing out that the relevant duty of care was to take reasonable steps to prevent violent assault, that each plaintiff was the victim of a violent assault, and that the damage sustained by the plaintiffs was "the very kind of thing" which the relevant duty obliged Adeels Palace to take reasonable steps to prevent... In the present case, in contrast, the "but for" test of factual causation was not established. It was not shown to be more probable than not that, but for the absence of security personnel (whether at the door or even on the floor of the restaurant), the shootings would not have taken place. That is, the absence of security personnel at Adeels Palace on the night the plaintiffs were shot was not a necessary condition of their being shot. Because the absence of security personnel was not a necessary condition of the occurrence of the harm to either plaintiff, s 5D(1) was not satisfied. (emphasis in original) 199Similarly, in my opinion it has not been shown by the appellant to be more probable than not that, but for the absence of the VZIG injection being given to the mother, the appellant would not have been afflicted by FVS. The highest that the case of the appellant in the present case can go, in light of the statistics, is that the giving of VZIG to the mother might have made the occurrence of FVS less likely and/or less severe. At [57] the Court in Adeels Palace went on to say: ...the submission that the plaintiffs' injuries in these cases were caused by the failure of Adeels Palace to take steps that might have made their occurrence less likely, should be rejected. 200In the present case, there was a very small percentage chance that the VZIG vaccine at the relevant time would have prevented the appellant from being afflicted or seriously affected by FVS. Given the very small size of the percentage in this case, no greater inference should be drawn than that the vaccine might have done something (being either to attenuate or prevent FVS). In Amaca Pty Ltd v Ellis (2010) 240 CLR 111 at [70], the Court said: ... no scientific or medical examination can now say, with certainty, what caused Mr Cotton's cancer or lung cancer in any other particular case. As explained at the outset of these reasons, despite this uncertainty, the courts must, and do, "reduce to legal certainty [a question] to which no other conclusive answer can be given" (69). The courts do that by asking whether it is more probable than not that X was a cause of Y. Saying only that exposure to asbestos may have been a cause of Mr Cotton's cancer is not a sufficient basis for attributing legal responsibility. Observing that a small percentage of cases of cancer were probably caused by exposure to asbestos does not identify whether an individual is one of that group [that would be a statistical likelihood measure, not a causal inference in the case at hand]. And given the small size of the percentage, the observation does not, without more, support the drawing of an inference in a particular case." (my emphasis) 201Similarly, but more to the point of any possible attenuative effect (as opposed to preventative effect) of the biologic drug, in Tabet v Gett, Hayne and Bell JJ (at [65]), agreeing specifically with Gummow A-CJ (at [46]), said: The respondent should not be held liable where what is said to have been lost was the possibility (as distinct from probability) that the brain damage suffered by the appellant would have been less severe than it was. (see also Kiefel J (at [101])). 202At most, not giving VZIG to the mother prevented the already existing very small statistical risk of the appellant contracting FVS from minutely decreasing (although I accept that in relative terms there was less than a 50% reduction in that very small risk). It did not cause the appellant's injury. If the effect of VZIG were simply to attenuate the effect of the disease, then all the administration of VZIG would have done is provide the possibility of an undefined better outcome. Given the quality (level III) of the Enders studies, they must be treated with caution. At best all the studies relied upon by the appellant did was to establish a possibility that the appellant would not have contracted (or would have had attenuated) FVS. In this case, that was a very small possibility. In Amaca v Booth (2011) 246 CLR 36, at [41] and [43], French CJ said: Causation in tort is not established merely because the allegedly tortious act or omission increased a risk of injury. The risk of an occurrence and the cause of the occurrence are quite different things. That proposition is obvious enough and not determinative of these appeals. The existence of an association or a positive statistical correlation between the occurrence of one event and the subsequent occurrence of another may be expressed as a possibility, which may be no greater than a "real chance" that, if the first event occurs, the second event will also occur. The mere existence of such an association or correlation does not justify a statement, relevant to factual causation in law, that the first event "creates" or "gives rise to" or "increases" the probability that the second event will occur. Such a statement contains an assumption that if the second event occurs it will have some causal connection to the first. However, if the association between two events is shown to have a causal explanation, then the conclusion may be open, if the second event should occur, that the first event has been at least a contributing cause of that occurrence. An after-the-event inference of causal connection may be reached on the civil standard of proof, namely, balance of probabilities, notwithstanding that the statistical correlation between the first event and the second event indicated, prospectively, no more than a "mere possibility" or "real chance" that the second event would occur given the first event. There may of course be cases in which the strength of the association, as measured by relative risk ratios, itself supports an inference of a causal connection." 203That there was (on the Enders table adjusted for the strike rate of the disease) a 39-44% (or alternatively a 45-48%) chance that the mother would be prevented from suffering the negative consequences of the disease (or the less than 1% chance for the foetus), addresses only whether the giving of VZIG might have prevented the injury to the foetus. The statistics do not support the conclusion that it is more likely than not that it would have prevented FVS in the foetus. Nor, in my opinion, does the evidence of the experts' clinical experience in the use of VZIG. The failure to administer VZIG in effect meant that the already small risk that the appellant would contract FVS if her mother contracted chickenpox was not reduced to an even smaller percentage. The statistical correlation between the two events was such that whether or not VZIG was administered there remained only a very small possibility that the appellant would contract FVS (and the statistical studies and empirical evidence did not support the conclusion that it was more likely than not that the administration of VZIG would have prevented the mother contracting varicella, which was what gave rise to the risk). 204Therefore, the failure to give the VZIG to the mother was not in my view established to have been a necessary condition of the occurrence of FVS in the appellant. Even if the drug had been given to the mother, it was more likely than not that the mother would have contracted varicella and there remained a risk that the appellant would have been afflicted by FVS. The failure to administer VZIG did not in my opinion materially contribute to that risk (particularly if the effect of administration of VZIG would have been only to attenuate the effect of the infection and not to prevent it). 205In the present case, the appellant relied on Chappel v Hart (1998) 195 CLR 232 (and, in particular, what McHugh J (in the minority) there said) as supporting a finding of causation. There, the majority of the Court found that the fact that the risk might be miniscule did not excuse the need to warn of the risk. However, in that case it was found that if the warning was given the surgery would not have gone ahead (and there was no doubt that the surgery had caused the injury). The negligent failure to warn in that sense was a necessary condition of that plaintiff undergoing the surgery which led to the damage the subject of the claim. It was not any negligence in the performance of the surgical procedure that caused the injury. 206Here, even if the warning had been given (and the drug administered), that would not have changed, on balance, the risk that the damage would have been suffered. 207In Adeels Palace, the High Court considered the passage in Chappel v Hart at [53], stating at [52]-[53]: Counsel for the plaintiffs, in this Court, relied upon passages in Chappel v Hart (53). But in that case the majority proceeded on the basis that but for the failure to warn the event would not have happened; the question then was whether certain additional factors, combined with the satisfaction of the "but for" test, were sufficient to establish causation. In the present case, in contrast, the "but for" test of factual causation was not established. It was not shown to be more probable than not that, but for the absence of security personnel (whether at the door or even on the floor of the restaurant), the shootings would not have taken place. That is, the absence of security personnel at Adeels Palace on the night the plaintiffs were shot was not a necessary condition of their being shot. 208The necessary factors in establishing a causal connection set out in the judgment of McHugh J in Chappel v Hart would also tend to the same conclusion (though here, for the reasons above, such factors are applied in a different factual matrix). His Honour said (at [34]): ... no causal connection will exist where the plaintiff suffered injury at some other place or some other time unless the change of place or time increased the risk of injury 209Insofar as the administration of VZIG might have changed (at a level that was less than 1%) the chance that the appellant might possibly have been prevented from contracting FVS or that its effect might have been attenuated, I note that in Chappel v Hart, McHugh J went on to say (also at [34]) that: ... no causal connection will exist if the eventuation of the risk is so statistically improbable as not to be fairly attributable to the defendant's omission. 210Basten JA has drawn attention to the lack of evidence on this issue (at [23]). There is not in my view a sufficient basis in the evidence before the Court from which to conclude that administration of VZIG at a dose sufficient only to attenuate the effects of the disease in the mother would lead to a material (or any) reduction in the risk of infection of foetuses. This is because the statistical value of the studies referred to by Professor Curtis is extremely limited by reason of the small cohorts and different dosages involved in those studies. In other words, the evidence says only that a VZIG vaccine given to the mother might possibly have had some unknown effect on the appellant. 211As adverted to earlier, a small reduction in the risk to the foetus of contracting FVS, in absolute terms, from 1-2% to somewhere below 1% may be differently described by reference to what that reduction amounts to in relative terms. If calculated in relative terms, the appellant's chances of contracting FVS would decrease, by the administration of VZIG to the mother, within a range of 39-48%. The absolute risk reduction to the foetus is less than 1% on any calculation. The risk to the foetus remains that it is more likely than not (on either description of the risk) that the vaccination of the mother would not have prevented the foetus contracting FVS. 212In my opinion, in the present case the appellant has not established that the respondent's breach of its duty caused or materially contributed to the harm of which complaint is made. 213I turn then to the possible application of s 5D(2) of the Civil Liability Act, pursuant to which, in exceptional cases, a court may award damages for negligence even where factual causation is not established. 214Section 5D(2) provides: In determining in an exceptional case, in accordance with established principles, whether negligence that cannot be established as a necessary condition of the occurrence of harm should be accepted as establishing factual causation, the court is to consider (amongst other relevant things) whether or not and why responsibility for the harm should be imposed on the negligent party. 215The Ipp Report (at [7.30]), which led to the s 5D(2) amendment, used Fairchild v Glenhaven Funeral Services Ltd [2003] 1 AC 32 as an example of a situation where an application under such a section would be appropriate. There, successive employers had exposed the plaintiff to asbestos dust but scientific evidence as to the aetiology of mesothelioma did not justify a conclusion in relation to any of the employers that but for the negligence of that employer the plaintiff would not have contracted the disease. The House of Lords found that each defendant's wrongdoing materially contributed to the harm suffered by the plaintiff. 216The other example used in the Ipp Report was that of Bonnington Castings v Wardlaw [1956] AC 613. In that case, an employee contracted pneumoconiosis from silica dust inhalation, from equipment in his workplace. The state of scientific knowledge at the time did not enable proof as to which of the dust from a pneumatic hammer (which clearly spread silica dust, but as there was no way to prevent this, gave rise to no breach of duty) and the dust from swing grinders (where an improper set-up of this equipment increased silica dust flow and was found to be in breach of duty) had caused the employee's pneumoconiosis. Thus it was not certain whether the breach regarding use of the swing grinder caused the disease. Lord Reid said "the real question is whether the swing grinders materially contributed to the disease" at 621. His Lordship concluded by stating that "it is proved not only that the swing grinders may well have contributed but that they did in fact contribute a quota of silica dust which was not negligible to the [worker's] lungs and therefore did help to produce the disease". 217The Ipp Committee stated, in relation to the principle applied in Bonnington (at [7.28]) that "[t]he effect of this rule is that a defendant may be liable for the total harm suffered by a plaintiff even though it cannot be said that, but for the conduct of the defendant, the plaintiff would not have suffered the total harm; and that it can only be said that but for the conduct of the defendant the plaintiff would not have suffered some of that harm" (emphasis in original). It was suggested by the Committee that a legislative provision resembling s 5D(2) should be enacted to bridge the evidentiary gap that had existed in Fairchild and in Bonnington. 218The Committee considered that in certain circumstances it might be appropriate to allow proof that negligent conduct materially contributed to harm or the risk of harm, to satisfy the requirement for proof of factual causation but noted that this was a normative rule dependent on a value judgment as to how costs of injury and death should be allocated. 219In Adeels Palace, at 443, the High Court said: It may be that s 5D(2) was enacted to deal with cases exemplified by the House of Lords decision in Fairchild v Glenhaven Funeral Services Ltd where plaintiffs suffering from mesothelioma had been exposed to asbestos in successive employments. Whether or how s 5D(2) would be engaged in such a case need not be decided now. The present cases are very different. No analogy can be drawn with cases like Fairchild. Rather, it would be contrary to established principles to hold Adeels Palace responsible in negligence if not providing security was not a necessary condition of the occurrence of the harm but providing security might have deterred or prevented its occurrence, or might have resulted in harm being suffered by someone other than, or in addition to, the plaintiffs... Accordingly, the submission that the plaintiffs' injuries in these cases were caused by the failure of Adeels Palace to take steps that might have made their occurrence less likely, should be rejected. 220In much the same way as Basten JA speaks of the result at first instance being counter-intuitive, there is at first blush attraction in an argument that this was an exceptional case where factual causation could not be proved, because of the ethical difficulties in conducting tests that would be more reliable in assessing the statistical link between administration of VZIG and the incidence of FVS. 221However, the appellant did not raise an argument based on s 5D(2) at trial and it did not form a ground of appeal in the Notice of Appeal (Red 298-299). Mr Romaniuk in oral submissions suggested that the appellant placed reliance upon s 5D(2) only if, applying McHugh J's analysis in Chappel v Hart, the Court nevertheless considered that causation under s 5D(1)(a) was not made out. However, as Counsel for the respondent (Mr Windsor) noted, exceptional circumstances were not pleaded. It was submitted by Mr Windsor that now to permit such a claim would give rise to procedural unfairness in that there might have been resource arguments that the respondent (a public authority) might then have raised that could have had an impact on causation and that other (unidentified) evidence might have been adduced. 222As the issue was not raised on the pleadings, it is not appropriate for it to be considered further. I simply note that the application of s 5D(2) in the present case would appear to be contrary to the reasoning in Adeels Palace, having regard to the view I have formed on the issue of causation. 223Accordingly, I agree with Hoeben JA that the appeal should be dismissed with costs.

DISCLAIMER - Every effort has been made to comply with suppression orders or statutory provisions prohibiting publication that may apply to this judgment or decision. The onus remains on any person using material in the judgment or decision to ensure that the intended use of that material does not breach any such order or provision. Further enquiries may be directed to the Registry of the Court or Tribunal in which it was generated. Decision last updated: 14 June 2013