Hughes v SDN Children's Services

LOWER COURT JURISDICTION : Supreme Court - Common Law Division LOWER COURT 20908/94 FILE NUMBER(S) : LOWER COURT Studdert J JUDICIAL OFFICER :

COUNSEL : Appellants: D J Higgs SC/J Clarke Respondents: J Poulos QC/S Torrington SOLICITORS : Appellants: Charlton Shearman Respondents: J Sharpe, solicitor for NRMA Workers Compensation (NSW) (No 2) Pty Ltd (formerly known as HIH Workers Compensation (NSW) Pty Ltd) CATCHWORDS : Negligence - Causation CASES CITED: Seltsam Pty Ltd v McGuiness (2000) 49 NSWLR 262 DECISION : Appeal allowed

• 3 - IN THE SUPREME COURT OF NEW SOUTH WALES COURT OF APPEAL

CA 40483/00 SC 20908/94 MEAGHER JA BEAZLEY JA GILES JA

Friday, 8 February 2002 LINDA HUGHES & ANOR v SDN CHILDREN'S SERVICES INC

FACTS The first appellant was pregnant with the second appellant (Jacob) whilst employed by the respondent as a childcare worker. Shortly after birth it became apparent that Jacob had significant disabilities including features of cerebral palsy, neural deafness and mental retardation. Jacob was referred to Dr Grigor a specialist paediatric in internal medicine who diagnosed Jacob's disabilities as being caused by cytomegalovirus (CMV), a herpes type virus, which, Dr Grigor considered had been contracted by Jacob in utero. At trial the respondents put in issue that Jacob's disabilities were caused by CMV.

Infection of the foetus can occur either through primary maternal CMV infection or maternal reactivation infection. It was common ground that for the appellants to succeed in their claims for damages that they needed to establish that the injury to Jacob was caused by a primary as opposed to reactivated infection. It was not known whether the first appellant had had the virus prior to her pregnancy.

At first instance Studdert J held that the respondent had breached its duty to warn the first appellant of the dangers of contracting CMV whilst pregnant. He also held that Jacob's disabilities were caused by CMV. However, he held on the balance of probabilities that the first appellant was a carrier of the virus at the time of her pregnancy and the injury to Jacob was caused by a reactivation of the infection, rather than primary infection of the first appellant during pregnancy. The appellants challenged this finding on appeal.

HELD per Beazley JA (Meagher JA and Giles JA agreeing) (i) The trial judge made erred in a number of important respects in his assessment of the evidence, and in particular in his finding that the appellants' principal expert witness made a concession in relation to Jacob's immunological status. That evidence was otherwise unchallenged.

(ii) The combination of facts and circumstances, and in particular the pattern of Jacob's disabilities (a matter which the trial judge appeared to overlook), the fact that the first appellant had flu like symptoms (typical of CMV infection) early in her pregnancy, the date of the foetal insult, which corresponded to a contraction of the virus early in the first appellant's pregnancy all pointed to Jacob's disabilities being caused by primary infection.

(iii) On the balance of probabilities Jacob's disabilities were caused by congenital primary CMV.

(iv) Judgment for the first appellant in the sum of $150,000 to take effect as from 30 May 2000 (being the date of judgment at first instance).

(v) Judgment for the second appellant in the sum of $4,500,000 to take effect as from 30 May 2000.

(vi) The respondent is to pay the appellants' costs of the hearing below and of the appeal, but is to have a certificate under the Suitors' Fund Act 1951 (NSW), if so qualified, in respect of the appeal.

IN THE SUPREME COURT OF NEW SOUTH WALES COURT OF APPEAL

CA 40483/00 SC 20908/94 MEAGHER JA BEAZLEY JA GILES JA

Friday, 8 February 2002 LINDA HUGHES & ANOR v SDN CHILDREN'S SERVICES INC

JUDGMENT 1 MEAGHER JA: I agree with Beazley JA. 2 BEAZLEY JA: This is an appeal from the judgment of Studdert J delivered on 30 May 2000 in which his Honour gave a verdict in favour of the respondent on the appellants' claims for damages arising out of severe disabilities suffered by the second appellant (Jacob), the first appellant's child, during the course of the first appellant's pregnancy. The first appellant's claim was for nervous shock consequent upon learning of the damage to Jacob. Jacob's claim was related to his injury and continuing severe disabilities. There is no issue on the appeal as to duty of care or breach. The issue is as to causation. 3 At first instance, the respondent had brought a cross-claim against its various insurers. Nothing from the cross-claim arises on the appeal. 4 The first appellant was a certificated child care worker and had always worked in that capacity both in Australia and overseas, since graduating from Newcastle TAFE in 1986. During one period working overseas, the first appellant worked at a pre-school in Ladbrooke Grove in London. The significance of that will become apparent later. She commenced employment with the respondent as a day care worker in October 1990 and continued in that capacity until about February 1993. In late November 1991, the first appellant's pregnancy was confirmed. She informed the respondent of her pregnancy in early December 1991. Up until about February 1992, the first appellant was working with 2 year old children. From February 1992, her duties were altered so that she was working with children under two. The first appellant took a period of maternity leave commencing mid June 1992. The second appellant was born on 6 August 1992. 5 The first appellant had an uneventful pregnancy except for some flu like symptoms on Christmas Eve and Christmas Day 1991 and there were no complications with Jacob's birth. However, in October 1992 the first appellant became concerned about Jacob's development. After consulting her general practitioner Dr Honeyman, Jacob was referred to Dr Grigor, a senior specialist in paediatric internal medicine. Dr Grigor saw Jacob for the first time on 4 November 1992. By January 1993 it was apparent that Jacob had significant disabilities including features of cerebral palsy, neural deafness and mental retardation. At that time, Dr Grigor was reasonably confident that Jacob's disabilities were caused by cytomegalovirus (CMV) which he had contracted in utero (congenital CMV). 6 CMV is a herpes type virus and is common among children, particularly in childcare centre environments. Infected children may transmit the virus to children and adults through objects and surfaces such as toys. The acquisition and excretion of the virus is age related, so that a child who is very young will excrete a larger quantity of the virus although the "amount of virus in all children less than 5 years is very high". Adults working with infants and young children may be infected by changing dirty nappies and assisting young children with toileting. It is thought that the danger of greater infection from very young children is because the virus is shed into nappies, which are then handled by the adult, whereas with children who are toilet trained, the virus is principally shed into the toilet. 7 The virus has an incubation period of between about two to four weeks. The virus in both adults and children is usually mild, often being asymptomatic. If symptoms are experienced they are mild flu like symptoms such as fever, chill, sore throat and enlarged glands. 8 Experts have described the risk of contracting the virus for women in day care centres as "particularly high" for those looking after under two year olds. In fact, the statistical evidence in the case was that 60 - 70% of all Australian women are seropositive and that women who work in child care centres had an 8% to 30% increased chance over the general population of contracting the virus. However, there were significant variances in these statistical studies. 9 Women who contract the virus are at risk of passing it on to their foetus. Infection of the foetus can occur in one of two ways, either following primary maternal CMV infection (that is where the mother is first infected during the pregnancy) or maternal reactivation/secondary infection (where the mother has had the virus in the past and it reactivates in pregnancy). 10 It was common ground that for the appellants to succeed in their claims for damages, they needed to establish that the injury to Jacob was caused by a primary as opposed to reactivated infection of the first appellant whilst working for the respondent. It was not known whether she had had the virus prior to her pregnancy. Her work history was thus important in determining whether she was a carrier of CMV (that is, whether she was seropositive) prior to becoming pregnant.

Trial Judge's Finding Regarding Duty of Care and Breach of Duty 11 His Honour was satisfied that the evidence was sufficient to draw an inference that: "by mid 1991 the [respondent] had such knowledge available to it of the risks of exposure of child care workers to the CMV virus as attracted a duty in the [respondent] to alert the [first appellant] to the risk in her work environment of CMV infection, not only to her, but to any child she might be carrying … that the duty extended to requiring that a warning be given that the CMV virus could damage any such unborn child." 12 His Honour was satisfied that whilst the respondent did give warnings regarding matters of hygiene, such as hand washing, no specific warning regarding the dangers of CMV was given to the first appellant. The duty to warn the first appellant arose before she became pregnant because "it was foreseeable that a young woman of child-bearing age might become pregnant". 13 His Honour further held that the transfer of the first appellant to work with children under two in January 1992, after she had become pregnant added to the risk of contracting CMV and was therefore a further act of negligence. 14 His Honour's findings in relation to the respondent's duty of care and breach of duty are not challenged by the respondent on appeal. The sole issue is whether the appellants had established that Jacob's disabilities were causally related to the respondent's negligence.

Trial Judge's Findings Regarding Causation 15 Studdert J considered that the first appellant needed to prove two matters in order to establish that Jacob's disabilities were caused by the respondent's negligence: "(i) [O]n the assumption that [Jacob] did acquire primary congenital CMV, that had the [first appellant] been given the appropriate warning she would probably have avoided the risk of such acquisition… (ii) Secondly, [that Jacob's] disabling condition was primary congenital CMV acquired by reason of the [respondent's] negligence." 16 In answer to (i), his Honour accepted that, on the balance of probabilities, the first appellant would have given up childcare work if she had been warned of the risk to her child. This finding was not challenged on appeal. 17 That left the question whether Jacob suffered congenital CMV due to a primary infection of the first appellant whilst pregnant. In seeking to determine that issue his Honour posed the following two questions: "(i) Is [Jacob's] condition caused by congenital CMV?

(ii) If so, was this acquired in consequence of work related primary infection of his mother during her pregnancy, in turn passed on to Jacob?" 18 His Honour answered the first question affirmatively and that too was ultimately accepted by the respondent on the appeal. 19 As I have already indicated, it was not known whether the first appellant was seropositive or seronegative before her pregnancy. However, his Honour, after commenting that he had "found it to be extremely difficult to reach a conclusion on the evidence … as to whether the injury to [Jacob] was due to primary or reactivated infection", concluded that he was unable to determine on the balance of probabilities that the disease was caused by primary infection. The correctness of this finding is the sole issue on the appeal. 20 It is instructive at this point to set out the findings based on the medical evidence that led his Honour to his conclusion. His Honour said: "197 I now draw together the following findings:

(i) that the occurrence of reactivated infection in pregnancy is common, happening in one in three to one in two women;

(ii) that the chance of reactivated infection in the mother leading to severe disabilities in the child however is very low, and very much less than the incidence of primary infection in the mother leading to severe abnormality in the child, as the statistics below indicate;

(iii) that reactivation rarely leads to foetal infection, there being only one chance in 100 of this happening …

(iv) that such reactivation leads to a clinically apparent disease (not necessarily severe) even less frequently, there being only one chance in 250 of this happening …

(v) that, whilst reactivated infection in the mother is much more common than primary infection in the mother, it is more probable than not that the majority of symptomatic infections in infancy occur after primary CMV infection in the mother during pregnancy;

(vi) that primary infection when it occurs in pregnancy is much more likely to result in foetal infection than is reactivated infection, the incidence being one in two cases of primary maternal infection;

(vii) that such occurrence leads to clinically apparent disease in 10 to 15% of cases.

198 On the other side of the scales must be weighed these findings:

(i) that the very nature of the [first appellant's] occupation made it highly likely, statistically, that [she] would have had CMV before her pregnancy commenced;

(ii) that primary infection occurs in mothers in pregnancy relatively rarely, in only three percent of seronegative women…

(iii) that the severity of the disease in [Jacob] does not exclude reactivated infection in the mother as its source." 21 His Honour concluded at para 199 and 200: "… the severity of the disease in Jacob does not exclude reactivated infection...it is for the [first appellant] to prove that her infection was primary. The statistically high probability that the [first appellant] would have had CMV before her pregnancy stands in her way… The inevitable consequence of the [appellants'] failure to discharge the burden of proof on this issue is that the claims of the [appellants] must fail…" 22 In reaching this conclusion, his Honour also said at para 199: "I have reflected upon the support expressed by Professor Isaacs [that Jacob's condition was due to primary maternal infection] … I am not persuaded that Professor Isaacs' opinion as to this makes adequate allowance for the statistically high probability that at [the first appellant's] age and with her working history [she] would have been seropositive prior to the commencement of her pregnancy." 23 These findings and his Honour's conclusion requires a review of the medical evidence and the evidence of Professor Isaacs and Dr Rawlinson in particular.

The Medical Evidence 24 At the trial, there was a significant dispute as to whether Jacob's disabilities were due to the CMV virus at all. Dr Grigor, the physician to whom Jacob had been referred when his disabilities first became apparent, had, by January 1993, reached a fairly firm conclusion that his disabilities were caused by CMV which he had contracted in utero. His opinion was reinforced by a CT scan in early 1993. Then, in early 1994, Professor Gibson, Head of the Children's Cochlear Implant Centre ordered an MRI scan which confirmed Dr Grigor's clinical diagnosis. Dr Grigor considered, on all the evidence, that the insult had occurred in the second trimester of pregnancy. 25 In the second half of 1993, Professor Isaacs, a clinical professor of Paediatric Infectious Diseases at the University of Sydney and who was, at the time, Head of the Department of Immunology and Infectious Diseases at the Children's Hospital was asked to provide a medico legal opinion as to Jacob's condition. In his first report of 9 September 1993 Dr Isaacs expressed the opinion that Jacob was suffering from the effects of congenital CMV. 26 In June 1999, Professor Isaacs was asked to report, inter alia, upon whether Jacob's disabilities were due to congenital CMV infection and whether that most likely was the result of primary or reactivated infection. He reported: "… Congenital CMV infection can follow primary maternal CMV infection … or maternal reactivation … Although severe congenital infection can occur following either primary or secondary maternal infection, it is far more likely to follow primary infection. …" 27 He concluded: "If Jacob Hughes did indeed have congenital CMV infection, the severity of his disease means that it is far more likely to have been a result of primary maternal CMV infection than reactivation." 28 During the course of his evidence at trial, Professor Isaacs said that the MRI and CT scans carried out on Jacob and the pattern of Jacob's abnormalities suggested that the first appellant had primary CMV infection 13-21 weeks post conception. 29 Dr Kendall, an English neuro-radiologist consultant specialising in paediatric neuro-radiology at the Royal Free Hospital, the Hospital for Sick Children and the National Hospital for Neurology and Neurosurgery also considered that Jacob's MRI and CT scans "strongly support[ed] the diagnosis of congenital [CMV] infection in the second trimester". In a second report he expressed surprise at the fact that the respondent's neurologist was questioning the diagnosis. 30 The respondent's expert was Dr Rawlinson, an expert infectious diseases physician and virologist and Associate Professor of Medicine at the Prince of Wales Hospital, which is part of the University of New South Wales. He has made a specialty study of CMV, including in the area of paediatrics. Dr Rawlinson diagnosed the appellant as suffering from cerebral palsy but concluded that his disabilities were not caused by congenital CMV. He accepted, however, that the MRI and CT scans indicated a neurological foetal insult, consistent with some form of infective process. Dr Rawlinson said that the time frame of the insult was outside his expertise. 31 At trial, Dr Rawlinson maintained his view that Jacob's disabilities were not caused by congenital CMV. 32 It was also the respondent's case at trial that if the appellants did establish that Jacob's disabilities were caused by congenital CMV, it was due to reactivated not primary infection. However, Dr Rawlinson did not support the respondent's case on this issue. In his report of 3 September 1998, he stated: "Intrauterine CMV infection results almost always from primary infection of the mother. Jacob Hughes suffers from severe disability, which never results from the rare instances of reactivation of infection or reinfection during pregnancy ." (emphasis added) 33 It followed on this view that if the appellants established that Jacob's condition was caused by congenital CMV it could only have been due to primary infection. 34 Professor Isaacs was not as categorical as Dr Rawlinson. His view, as I have indicated above, was that severe abnormalities could occur from reactivated CMV but that such occurrences were rare. 35 As part of its approach in meeting the appellants' case that this was a case of primary infection, the respondent placed significant reliance on the high statistical possibility that the first appellant was seropositive prior to her pregnancy. This became a feature of the cross examination of Professor Isaacs in particular. 36 In his responses in cross-examination, Professor Isaacs stated that he had always approached his consideration of the matter on the basis that the first appellant had a relatively high statistical possibility of being seropositive prior to Jacob's conception. In this regard he acknowledged that it was also relevant to take into account the period and circumstances of her employment in child care. He recognised that each time she worked with 'under two's' the risk of contracting CMV was increased. 37 Professor Isaacs had not been aware of the first appellant's employment at the preschool at Ladbrooke Grove, a lower socio economic area in London with a migrant population (especially West Indian), where it might have been expected that she had a higher risk of contracting CMV. Professor Isaacs accepted that such matters would be relevant "in a general sense" to the assessment of whether Jacob's abnormalities were due to primary or reactivated CMV. 38 Professor Isaacs agreed in cross examination that the range of statistical possibility of the first appellant being seropositive was in the order of 80%-90%. He said that was the statistical possibility that he had taken into account in arriving at his conclusion that Jacob's condition was due to primary infection. With one exception arising out of material in a study conducted out of the University of Alabama, to which I will refer later, it was never suggested to Professor Isaacs, nor by any of the other experts, that the statistical possibilities were higher than 80-90%. Dr Rawlinson stated the likelihood of the first appellant being seropositive was 85-90%. 39 Professor Isaacs also expressed the view that if the first appellant was seronegative at the commencement of her pregnancy, she had a 10 fold increased risk of contracting CMV in the workplace setting she was in during pregnancy. This evidence was not contradicted and was in accord with the general statistical evidence in the case. His Honour appears to have accepted this evidence at least in a general sense: see para 13 above. 40 It was suggested to Professor Isaacs that the statistical possibilities were also relevant when making a clinical judgment. Professor Isaacs responded: "… the question I am being asked initially is, did I think Jacob has CMV? Yes, I do. Was it likely to have been primary? Yes, because that is the great strength of the evidence of the exact type and so on. That is a more complicated issue and then all sorts of other issues that you are bringing up come into play. But it doesn't materially alter the question, does Jacob have CMV effects? And it is likely to be primary or second[ary]. However likely that [the first appellant] was positive for CMV before this started, it still makes it highly unlikely this is secondary or reactivation CMV. It is likely to be primary ." (emphasis added)

The Alabama Study 41 After the appellants' experts had completed their evidence and during the course of Dr Rawlinson's evidence, the respondent introduced a study by Boppana et al from the University of Alabama (the Alabama study) dated July 1999. The Alabama study was conducted between 1991 and 1997. Its subjects included 246 children with congenital CMV infection. Forty-seven of these children had clinical abnormalities following birth and of those infants, eight were born to mothers with confirmed reactivated infection. Another eight cases were cases of abnormalities following primary maternal infection. 42 The study findings were that the range of severity of clinical abnormalities during the newborn period did not differ in the cases determined to be primary infection cases and those that were not. In the paper prepared following the study the authors noted that: "Based on [previous] findings, it has been believed widely that symptomatic congenital CMV infection almost always occurs after a primary maternal infection. However, the results of the present study clearly document that symptomatic congenital CMV infection after a recurrent maternal infection occurs more frequently than has been thought previously …" 43 The concluding paragraph of the report contained the following reservation: "A major limitation of the present study is our inability to categorize the type of maternal infection in more than half of the children with symptomatic congenital CMV infection born during the study period. It is possible, and perhaps likely, that the vast majority of symptomatic infections occur after a primary CMV infection during pregnancy. It is also possible that categorization of most maternal infections will lead to identification of an increased proportion of children who were born to mothers with preexisting immunity among infants with symptomatic congenital CMV infection …" (emphasis added) 44 Dr Rawlinson was examined in relation to the study and Professor Isaacs and Dr Grigor were recalled. Dr Kendall was not recalled after the introduction into evidence of the Alabama study, although his Honour records that the "opportunity was given" for his recall. No other comment was made in that regard and no Jones v Dunkel inference was drawn. This is not surprising as Dr Kendall had given evidence that it had been recognised for about ten years that severe abnormalities could result from reactivated infection. Dr Kendall never purported to provide an opinion about whether Jacob's disabilities were caused by primary as opposed to reactivated infection, as it was outside his area of expertise. 45 Perhaps it should be stated that notwithstanding the significance which this study seemed to assume in the respondent's case at trial, its ultimate significance for this case was limited. The study appeared to be a reversal of an earlier study by the same authors where the result had been that nearly all cases of severe congenital disabilities were due to primary infection. This study seemed to indicate that the percentage of severe disabilities due to reactivated infection was higher than previously thought. However, for the reasons I indicate below, the results of the study, had little, if any, effect on the views of the medical experts except for that of Dr Rawlinson. The effect on his evidence was major. He was asked: "was there … an outcome [from the study] that the damage suffered by the foetus in reactivation could be just as severe as in primary CMV?" 46 He responded: "Yes, that's again different to conventional wisdom." 47 Later, he was asked by his Honour: "Q Is that your understanding of the overall result of the survey, that the result in terms of severity didn't differ much? A Yes. Q Between the recurrent infection group and the primary infection group? A Yes, and that was what was different about the study is to what our conventional wisdom was …" 48 He thus said that if he was writing his report of 3 September 1998 (see paras 32 and 33 above) after considering the Alabama study he would have expressed the view that Jacob's severe disability could have resulted from either primary or reactivated infection and that was not possible to distinguish between the two on clinical grounds. 49 The postulated outcome was not so startling for Professor Isaacs. When Professor Isaacs was asked whether the Alabama study changed his views he said that the study did not alter the balance of probabilities in Jacob's case and that the "most likely cause [of Jacob's CMV was] primary … maternal infection". As he pointed out, he had always been of the opinion that severe disability could be caused by reactivated infection. He was asked: "Q What I want to be clear on, what is your view now, having regard to this study, as to whether you can categorise the type of infection, by reference to the severity of the infection in the child?

A No, I don't believe you can. I never did believe you could." (emphasis added) 50 The trial judge observed that the last response did not seem to rest altogether comfortably with Professor Isaacs' report of 27 June 1999, see paras 26 and 27 above. I do not see any inconsistency. In his report Professor Isaacs was speaking of probabilities. The question to which he responded was concerned with categorisation of a type of infection. 51 Professor Isaacs considered that the real importance of the study was the possibility that seropositive mothers might have acquired a new strain of CMV. He was asked: "Q As I understand the study, the likelihood in the study, although there is a possibility of reinfection, the likelihood is that they were dealing with a reactivation rather than a [primary infection]? 52 He explained: "A No, I don't think they know that at all. They are wondering what is going on and with modern science, if you could get the old strain of CMV and then the one in the baby, they would be able to look at and compare them. Unfortunately they haven't got the old strain, so that's going to be their next study, to see whether there is reactivation of Mum's own CMV, or infection with the completely new strain. And I think it's probably more likely, that it is infection with a new strain . We know that does occur in other situations, for example, in transplant patients. They get a new strain of CMV but I think that it's an exciting new prospect in the study." (emphasis added) 53 Professor Isaacs added that if it was a new strain of CMV, that brought up exactly the same problems as primary infection. 54 Senior counsel for the cross respondent attempted to use the study for a further purpose, namely to establish that the first appellant was in the same statistical class as the subjects of the study. The basis for this approach was the first appellant's employment in child care centres and in particular her employment at Ladbrooke Grove. Professor Isaacs disagreed, pointing out: "A … you look at the population as a whole, so that a very young black single mother in the Alabama population, would perhaps have a 99 percent chance of being CMV positive.

Interestingly, senior counsel for the cross-respondent did not seek to press such a view on Dr Rawlinson.

55 Dr Grigor expressed similar views as Professor Isaacs as to the significance of the study. Like Dr Rawlinson, he was not asked whether the study meant that the first appellant was in a higher statistical class. 56 Notwithstanding the use which the respondents sought to make of the Alabama study, his Honour found that "the majority of symptomatic infections in infancy occur after primary CMV infection in the mother during pregnancy". He added: "What the study indicates is that the effects of the reactivation virus can be just as severe as the effects from a primary infection. This feature of the study appears to have influenced [Dr] Rawlinson to change his mind in the respects indicated, but Professor Isaacs had considered before he saw the report that severe congenital infection could follow reactivated maternal infection (indeed he indicated this in his report of 27 June 1999), and the Alabama study has not altered his opinion." (emphasis added) 57 All experts agreed that a diagnosis of congenital CMV was best made by a testing of the cord blood shortly after the birth. However, that opportunity had been missed in Jacob's case.

Significance of Results of the Antibody Test at 3 Months and 4½ Months 58 As stated earlier, Dr Grigor was the first specialist to have the clinical care of Jacob. He suspected that congenital CMV was the cause of Jacob's disabilities. On 4 November 1992 he arranged for Jacob to have a urine test and blood test. As part of the blood test, he ordered a TORCH study, which is a study looking for five specific infections including CMV. The test for CMV within the TORCH spectrum is known as a complement fixation test (CFT). That test measures a combination of IgM and IgG. The urine test was positive for CMV. Negative results were reported for all infections within the TORCH spectrum. Dr Grigor found this surprising because he suspected that Jacob had congenital CMV. Dr Grigor thus requested a more specific test to detect if there were any detectable M and G immuglobins (IgM and IgG respectively). 59 IgM and IgG are both antibodies produced at the time the body suffers an infection. IgM is produced at the time of the infection, its function being to 'mop up' the virus. If found, it indicates recent infection. IgM disappears from the person's system within a few weeks and does not cross the placenta. IgG is produced 2-3 weeks after an infection and persists in the body. Its function is to assist the body to resist further infection but it is relatively ineffective in the case of CMV. 60 The further testing requested by Dr Grigor was not done at that time. 61 Dr Grigor ordered a further blood test on Jacob six weeks later, on 18 December 1992. It too returned a negative result, which Dr Grigor again found surprising. However, the negative result did not cause him to change his provisional diagnosis of congenital CMV. 62 Dr Grigor said that when trying to reconcile the two negative results with his opinion: "I reasoned that it was most unlikely that the infection was peri or postnatal because I would have expected a baby to produce antibodies in that period of time in a normal way and one would see a change between the level of the complement CFT at three months and one at three months plus six weeks. … Whereas if it had occurred in utero, … I could reason, without being aware that it was proven to be so in other circumstances, other cases, that the baby had failed to manufacture antibodies in utero, on the basis that, that his body recognised the [CMV] as itself and so did not see itself as a foreign antigen, a foreign stimulant to producing antibodies." 63 It was suggested by senior counsel for the cross-respondent at trial that there was in fact an alternative explanation for the negative results, namely that Jacob was developing CMV between the two tests. Dr Grigor rejected that explanation because of the positive urine test result at three months. 64 Dr Grigor believed he again asked for an IgM/IgG specific test in December 1992 but that the laboratory advised that there was no point in its being carried out. There was some confusion about this part of the evidence. The better view seems to be that the more specific testing was not carried out. 65 Professor Isaacs also expressed the view that it was surprising that if Jacob did have congenital CMV he had no detectable antibodies at 3 months and at 4½ months. He explained: "If you then showed me the antibody results, I would say, 'This is very strange. I don't understand why there is no antibody in Jacob's blood, both at three months and possibly that traced at four and a half months. It doesn't make sense.' If you try to ask me to explain why that would be, my best explanation for that is that he's tolerant and that would only happen following primary maternal infection. In other words, he is incapable of producing antibodies." 66 Professor Isaacs' reference to "possibly … four and a half months" in this passage related to the confusion as to whether the more sensitive testing ordered by Dr Grigor was in fact carried out. When that matter was clarified, Professor Isaacs was of the view that Jacob's antibody count at three months was only explicable by primary infection. He first explained: "Q You said that if it was, if there had been a reactivation, … you would have expected lots and lots of antibodies? A That's correct. Q In whom? A Both the mother and the baby, because the antibody goes across the placenta. Q That would be likely, whether it was primary or secondary infection wouldn't it? A No. If I have to explain it - we now assume that Jacob did have symptomatic congenital CMV infection. Now as a doctor if I am asked, 'Do you think this was primary or non-primary?', I would say, 'On the basis of the results, of the probabilities, it's likely to be primary but not absolute', as his Honour says. If you then showed me the antibody results, I would say, 'This is very strange. I don't understand why there is no antibody in Jacob's blood, both at three months and possibly that traced at four and a half months. It doesn't make sense.' If you try to ask me to explain why that would be, my best explanation for that is that he's tolerant and that would only happen following primary maternal infection. In other words, he is incapable of producing antibodies. … Q … the same antibody created IgG will show in the foetus, whether or not it is a primary or secondary infection, won't it? A There are two points I would like to make to that. The first is it won't - it doesn't usually have the same effect on the foetus, because the antibody that is around already, is thought to be relatively protective to the foetus. Having said that, if you get a reinfection that's, for example, if you have had measles before and you get measles again, you produce a vast amount of antibodies. It is a question of degree. You would [expect] lots of antibodies with a reinfection and not so many antibodies necessarily with a primary one." 67 A little later he said: "A. … there are three possibilities. One is that Jacob did not have congenital CMV infection. That's why he has no antibody but it's all coincidence that he caught CMV incidentally after birth and all his malformations is a co-incidence. They are due to something else. If we reject that possibility, there are two major considerations left. One is that he caught CMV infection inutero, as a result of primary maternal CMV and the other is as a result of non-primary. If we assume it's one of those, we then have to explain why he does not produce antibodies at three months in both settings. The books would tell us we would expect to find antibodies at three months, a lot of it spilling over from his mother and that as that waned he would produce his own antibodies. So at three months and again at four and a half months, we would expect high levels of IgG. The best explanation I can come up with, to explain this unlikely phenomenon, is that Jacob has tolerance. That is much more likely to occur. In fact, I think it could only occur with primary maternal CMV, not non-primary. HIS HONOUR: Q. Why? … A. Because he, as a baby, would have needed to - it has to be a new antigen and if it was non-primary, he would effectively have already seen the virus and therefore could not become tolerant to it. Tolerance is a phenomenon whereby it's the very first time that a virus gets into your system." 68 The absence of antibodies at 3 and 4½ months was relevant to the respondent's case, including Dr Rawlinson's opinion, that Jacob did not have congenital CMV. Dr Rawlinson gave evidence that if Jacob had congenital CMV (whether primary or reactivated) there was an 85-90% probability of a positive IgG at three months. He explained the positive urine test at 3 months as indicating that Jacob had acquired CMV post-natally and probably at about 3 months. In that case, the only explanation he could give for Jacob not having antibodies at 4½ months was that he had an immature immune system. Neither Professor Isaacs or Dr Grigor accepted this. 69 Dr Rawlinson considered that Dr Grigor and Professor Isaacs' explanation as to why Jacob did not have antibodies at 3 or 4½ months, assuming he had congenital CMV to be a plausible hypothesis but one that he did not accept. He remained of the opinion that the reason for the absence of antibodies was because Jacob did not have congenital CMV. 70 It is of importance in the appeal, for reasons which will become apparent, that the trial judge said that he did "not find that evidence that the absence of antibodies ruled out reactivated infection to be persuasive". It is to be noted that evidence to that effect was not included in the catalogue of findings drawn together by his Honour (see para 20 above). It is necessary to refer to the particular evidence upon which his Honour relied in reaching this conclusion. 71 During the course of his cross examination, Professor Isaacs was being questioned in relation to the antibody evidence. He was asked a series of questions by his Honour. Immediately after his Honour's questioning referred to at para 66 above, senior counsel for the respondent asked him this question: "Q As a matter of logic, the reason why Jacob was injured, was because he had a viral attack, on your thesis. Is that right?

Q So that is the first time when he has seen the virus, on your thesis. Is that right?

Q So whether or not that virus is a result of mother's reactivation virus, or alternatively a primary infection, it matters not a wit [sic: whit], in terms of your thesis, does it, because for Jacob to be infected and allowing some credit to this study in paediatrics, that the damage can be just as severe with reactivation, if it is the first time he sees it. It doesn't matter a wit [sic] whether it's primary or secondary, because it was the first time that he comes into contact with the virus that is important, not when the mother comes into contact with the virus?

Trial Judge's Critical Finding in Respect of Professor Isaacs' Evidence 72 In coming to his conclusion that the evidence in relation to antibodies was not persuasive, his Honour considered what he referred to as "the concession in the last answer above" to be relevant. With respect to his Honour, I consider he either misunderstood this answer or placed too much reliance on it. The question was long and not very specific. The question, by its express terms related to Professor Isaacs' "thesis" that Jacob was injured because of a viral infection. The question postulated that "the damage can be just as severe" with a secondary infection because the foetus is responding to the virus for the first time. In its terms, the question did not refer to the presence or absence of antibodies, which, on the medical evidence in the case is a very specific issue. Professor Isaacs responded to the question asked. He was tentatively agreeing with the cross-examiner's rationale for the acknowledged possibility of severe damage from secondary infection. 73 In my opinion, his Honour fell into error in his understanding of the nature or extent of the 'concession' made by Professor Isaacs. 74 Having found Professor Isaacs evidence as to antibodies to be unpersuasive, the trial judge said that, as he understood Professor Isaacs' evidence in its entirety, including his response to the Alabama study, it was not the severity of Jacob's disabilities which led him to the view that Jacob's disabilities were caused by primary rather than reactivated infection. This is correct, in that Professor Isaacs' view was founded on more than the severity of Jacob's disabilities: however, on my reading of Professor Isaacs' evidence the severity of Jacob's disabilities was one element in his coming to his view. The catalogue of findings drawn together by his Honour did include statistical regard to the severity of the condition. But his Honour seems to have failed to give effect to the severity of Jacob's disabilities as an element in Professor Isaacs' opinion when reflecting on the support expressed by Professor Isaacs for Jacob's condition being due to primary maternal infection. That is so because his Honour considered that Professor Isaacs had arrived at his conclusion that Jacob's disabilities were caused by primary infection not because of the severity of the disabilities but because there was much less chance of the foetus becoming symptomatically infected by a process of reactivation. His Honour reiterated this view in his conclusion at para 199 to which I have referred in para 22 of these reasons, namely, that he was not persuaded that Professor Isaacs had made adequate allowance for the first appellant's statistically high possibility of being seropositive at the commencement of her pregnancy. 75 Quite apart from his evidence concerning antibodies, in my opinion this does not correctly assess Professor Isaacs' evidence. First, Professor Isaacs did attribute significance to the severity of Jacob's disabilities. Secondly, he gave evidence as to the status of the statistical possibilities and its relevance see paras 36 to 38 above, and that evidence was in accord with statistical possibilities advanced by the respondent. It is difficult to understand, therefore, why his Honour considered that Professor Isaacs did not give sufficient weight to this issue. 76 And Professor Isaacs' evidence went beyond the statistical evidence. I have already referred to this in some detail, and the following evidence emphasises Professor Isaacs' point. 77 During the course of further cross-examination after the introduction of the Alabama study, he said: "… I am not saying [the authors of the Alabama study] are right or wrong … I am taking their information and trying to address it to the current case. … But in our current case, it's only of marginal impact and why I say that it's only of marginal impact, is that if we try and address what happened to Jacob and we now say, we are almost certain he did have symptomatic congenital CMV, we have to say, is it more likely that it was due to primary maternal infection, or due to non-primary and if, as I believe, the rate of non-primary infection here is much lower, it's still most likely to be due to primary infection. When we add in the antibody results on Jacob, taken at three months, if this was non-primary, they really ought to be sky-high . He should have had massive IgG antibodies here, because his mother would have, because his mother would have had a reinfection … and would have produced masses of antibody. So I still say that by far the most likely scenario here, is that Jacob was infected by primary maternal CMV infection." (emphasis added) 78 In my opinion, the trial judge's conclusion fails to give proper weight to Professor Isaacs' clinical assessment of the cause of Jacob's CMV, given that Professor Isaacs had appropriate regard to the first appellant's statistical possibility of being seropositive prior to conception. For my part I consider that was a critical oversight and amounts to error. I will return to this matter shortly.

Trial Judge's Reliance on Statistical Evidence 79 The approach which his Honour took in determining the case, having found Professor Isaacs' evidence on the antibodies unpersuasive, was to return to the statistical evidence. He said at para 191: "I find on the evidence that the incidence of reactivation of CMV in the mother causing severe abnormality in the infant is very much less than the incidence of primary infection in the mother resulting in severe abnormality in the infant."

80 At para 193, he further observed other evidence of Dr Rawlinson: "…emphasises, statistically, how rare is the chance of reactivated infection leading to clinically apparent disease in the child. It is as [Dr] Rawlinson agreed 'an extraordinarily small prospect'". 81 This finding became encompassed in his Honour's ultimate findings at paras 197 and 198 which are set out earlier. The critical ones for present purposes are those made in para 198, because it was due to those considerations that his Honour found that it was the high statistical probability that the first appellant was seropositive before her pregnancy which "stands in her way".

Submissions on Appeal 82 The appellants approached the appeal on the basis that his Honour had incorrectly assessed the statistical evidence relating to the probabilities of the maternal infection being as a result of reactivation rather than primary infection. They stated in para 5 of their written submissions: "The [appellants'] appeal proceeds on an analysis of the statistical evidence adduced and a comparison between the relative likelihoods of [Jacob's] infection and disease being caused by either a reactivated infection or a primary infection in the [first appellant] during pregnancy." 83 Specifically, they contended that the trial judge failed properly to interpret and utilise Dr Rawlinson's figures. They proceeded to set out, through a statistical formula, the chance of the second appellant having suffered injury as a result of primary infection as opposed to reactivated infection. Using different seropositive rates they specified the statistical results. It is clear that the appellants' counsel relied on his Honour's findings at paras 197 and 198 to justify the figures used in the formula. 84 On the appellants' statistical analysis they contended: "Given the finding of fact that his Honour made, accepting as he did the statistics that we have transposed to these tables, and even assuming a worst case scenario against us ... [the figures] show that there is a minimum of an 18 fold greater prospect of the clinically apparent disease [in] the child being due to primary as opposed to reactivation infection." 85 The appellants also contended that the trial judge had erred in finding that there was a 'high likelihood' that the first appellant seroconverted before her pregnancy. After discussing the relevant evidence they stated that the more appropriate probability of the first appellant having seroconverted was 50%. At its highest the appellants argued that the rate would be no greater than 70%. The difficulty with that submission of course is that the appellants' own expert, Professor Isaacs, accepted that there was an 80 to 90% possibility. 86 The appellants also called into question Dr Rawlinson's figures in relation to the rate of primary infection during pregnancy. The appellants stated that the rate was likely to be higher given that the first appellant continued to work in a childcare centre during her pregnancy. Detailed figures were then provided to the Court as to what the rate of primary infection was. 87 The respondent submitted that the statistical formula and analysis relied on by the appellants was not put to the witnesses or the trial judge and therefore could not form the basis of their appeal. But in any event, the respondent's main submission was that there was ample statistical evidence to prove that it was highly likely that the first appellant was seropositive prior to becoming pregnant. According to the respondent's counsel, once this was accepted the appellants' appeal was defeated. In other words, once it was established that it was more probable than not that the first appellant was seropositive prior to her pregnancy, there was no need to consider the statistical evidence relating to the likelihood of defects occurring from primary or reactivation infection. The respondent also contended that the analysis by the appellants supported the trial judge's finding that there was a high likelihood that the first appellant was seropositive prior to pregnancy. 88 Whilst the respondent acknowledged that there was only a small prospect of clinically apparent disease to the foetus from reactivation, they contended that once there was a possibility that reactivation infection could cause severe neurological defect then it was open to his Honour to find that Jacob suffered injury as a result of reactivation. In particular, the respondent submitted: "In this matter [the first appellant's] past work history, her exposure to young children over a period of years together with the statistical evidence available from various articles is…powerful circumstantial evidence to establish the link between a reactivated CMV and the development of [Jacob's] neurological deficits." 89 Both parties relied upon Seltsam Pty Ltd v McGuiness (2000) 49 NSWLR 262 in support of their respective positions. 90 In Seltsam the plaintiff had renal cell carcinoma which he alleged had been caused by the inhalation of asbestos dust. He succeeded at first instance but, on appeal, it was held (by majority) that the plaintiff had not established the necessary causal link between his disease and the inhalation of asbestos. The case was conducted and decided principally on the basis of epidemiological evidence. 91 Spigelman CJ, in reviewing the epidemiological evidence put forward to establish causation both generally and in that case stated at 274: "Epidemiology is … concerned with the study of disease in human populations. It is not, of itself, directed to the circumstances of an individual case. For the purpose of determining whether exposure to a particular substance is the legal cause of a particular disease, epidemiology only provides evidence of possibility." 92 He explained that evidence of a possibility is admissible circumstantial evidence to be weighed with such other evidence as there is in determining whether on the balance of probabilities an inference of causation could or should be drawn in a particular case. He pointed out that if the evidence as a whole does not rise above a possibility, causation in law has not been proved. 93 This case is quite different from Seltsam. The legal task which confronted the trial judge was to determine whether the appellants had proved on the balance of probabilities that Jacob's disabilities were caused by primary infection in the first appellant during the pregnancy. The essential fact finding task necessary for that legal determination was whether it had been proved that the first appellant had a primary infection during the pregnancy.

Conclusions on Appeal 94 In my opinion, the appellants' statistical re-jigging, even if correct, does not assist the resolution of the matter for the reasons advanced by the respondent. The fact remains, the first appellants' statistical chances of being seropositive at the commencement of her pregnancy were high. Jacob's statistical chances of having congenital CMV due to primary infection were conversely extremely low. Because of the findings which he made in respect of Professor Isaacs evidence, his Honour was left with little more than a statistical case upon which to make his determination. On that case, I would agree with his Honour's conclusion. However, because I consider his Honour erred in the respects I have discussed, the question remains whether the appellants had established their case, on the balance of probabilities. 95 In my opinion, the following matters are relevant to the determination of that question. 96 First, the first appellant suffered flu like symptoms on Christmas Eve and Christmas Day. It is not known what the cause of the symptoms was. Professor Isaacs was of the view that the first appellant's flu-like symptoms over Christmas could have involved seroconversion. Support for contraction of the virus at this time is also to be found in the known incubation period of the virus and the apparent time of foetal insult (13-20 weeks). 97 Secondly, during her pregnancy, including the time in which she felt the flu like symptoms, the first appellant was working in an environment in which she was at a high risk of contracting the virus. Professor Isaacs said she had a 10 fold risk of contracting the virus in those circumstances if she was seronegative. No other figure was suggested or proved in the case. 98 Thirdly, although Professor Isaacs acknowledged that severe disabilities could be caused by reactivated CMV (he always having been of that opinion), he consistently regarded the severity of Jacob's disabilities as pointing to primary infection. That had been Dr Rawlinson's initial view, categorically held, although he changed it in the light of the Alabama study; for the reasons given by Professor Isaacs, the Alabama study may not have called for such a radical change. As well, Professor Isaacs consistently referred in his evidence to the clinical picture presented by the pattern of Jacob's abnormalities, the "exact type" of his disabilities as indicating primary as opposed to reactivated infection. Professor Isaacs was never cross-examined on this aspect of his evidence. No contrary evidence was called. In my opinion, there is a difference between "severity" and "exact type" and/or "pattern". This evidence appears to have been overlooked by his Honour. 99 Fourthly, both Professor Isaacs and Dr Grigor provided an explanation as to the absence of antibodies in Jacob's TORCH study. Dr Rawlinson considered this explanation to be plausible, if his own, virtually categorical view that the absence of antibodies indicated this was not a case of congenital CMV, was not correct. His Honour did not accept Dr Rawlinson's opinion that this was not congenital CMV. And if, as I think was the case, his Honour misunderstood the nature of the 'concession' made by Professor Isaacs, then his evidence that the absence of antibodies could only be explained by primary infection was unchallenged. Significantly, Dr Rawlinson was never asked by the respondent's counsel to express an opinion on this particular issue. The evidence had the benefit of cogency, given that in the case of reactivated infection, the mother would already be carrying IgG antibodies which she would pass on to the foetus. 100 Finally, Professor Isaacs paid due regard to statistical possibilities. When there are melded with the above clinical features, the statistical possibilities established in the case, notwithstanding what the statistics say about possibility I am of the opinion that the appellants established on the balance of probabilities that in this particular case Jacob's disabilities were caused by congenital primary CMV. The statistical chance that the first appellant was seronegative at the time of her pregnancy may have been low, but on the evidence she was and she contracted a primary infection in the early stages of her pregnancy. 101 The matter proceeded before Studdert J on the issue of liability only. His Honour noted that the parties had come to an agreement on damages, namely $150,000 to the first appellant and $4,500,000 to Jacob, plus costs. 102 I would propose the following orders: (i) Appeal allowed;

(iv) Judgment for the first appellant in the sum of $150,000 to take effect as from 30 May 2000 (being the date of judgment at first instance).

(v) Judgment for the second appellant in the sum of $4,500,000 to take effect as from 30 May 2000.

(vi) The respondent is to pay the appellants' costs of the hearing below and of the appeal, but is to have a certificate under the Suitors' Fund Act 1951 (NSW), if so qualified, in respect of the appeal. 103 GILES JA: I agree with Beazley JA.

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.