Dhupar v Lee

BRERETON JA: On 26 August 2014, the appellant Dr Nita Dhupar (“the doctor”), an obstetrician and gynaecologist, performed on the respondent, who has been given the pseudonym Jodie Lee (“the patient”), a laparoscopic tubal ligation [1] procedure utilising Filshie clips to occlude the fallopian tubes. [2] Notwithstanding the procedure, the patient subsequently fell pregnant and had a child delivered by caesarean section on 1 March 2016; inferentially she fell pregnant in late May 2015, about nine months after the procedure.

The patient sued the doctor in the District Court, contending that the doctor had been negligent in applying the left Filshie clip to a location and in a manner which did not achieve complete occlusion of the left fallopian tube, and in failing to undertake proper intraoperative inspection to ensure that the clip had been correctly applied. After a trial over ten sitting days between April and August 2019, Levy DCJ on 19 November 2020 gave judgment for the patient, for $408,700 plus costs. [3] The damages included damages for non-economic loss (psychological disorder, trauma of surgery etc), past economic loss and future loss of earning capacity, and past out-of-pocket expenses and future treatment expenses. The doctor appeals, both as to liability and as to damages.

The tubal ligation procedure in this case involved the application of two Filshie clips, one to each fallopian tube.

The following description of the application and operation of Filshie clips is largely drawn from the trial judge’s detailed account, [4] which was not contentious.

Filshie clips, a pair of which appears in the below photograph, are patented medical devices made of titanium with a silastic insert, designed to prevent conception by occluding the fallopian tubes.

In its open state (as in the above photograph), a Filshie clip has two articulated component jaws or arms that are connected at the base by a concealed riveted hinge. The upper jaw has a convex curve which, at a point a few millimetres from the end, turns into an upturned open concave curled lip. The lower jaw, which forms the base, is straight, but it incorporates a half return lip or locking latch, about 2mm in height, designed to receive the upturned or convex curled lip of the upper jaw during the locking process following initial apposition and closure. Before closure, the open portion of that latch faces towards the internal hinged portion. Both jaws have a thick clear flexible silastic lining, designed to squeeze and apply pressure to the fallopian tube when the clip is closed and locked. Once closed, the dimensions of a clip are about 14mm in length and about 3.5mm in width.

The laparoscopic application of a Filshie clip to a fallopian tube is a touch-free process, commencing with the clip being loaded into a pistol-like Sterishot applicator device. Mechanical pressure is then applied to a trigger on the applicator handle, which transmits pressure onto the concave curve of the upper jaw of the clip, which then progressively straightens and flattens as increasing pressure is applied through the applicator. The curve of the upper jaw of the clip then becomes locked into place as further and continued mechanical pressure is applied to push and advance the curled metal lip at the end of the upper jaw further forward, so that it engages with and locks under the return lip or latch of the lower jaw. Once a Filshie clip is locked, it becomes permanently closed and cannot be opened. There was undisputed evidence that the mechanical action of the applicator of the type used by the doctor in the procedure performed on the patient is inaudible and does not produce any “click” sound.

The following two diagrams provide a broad overview of the process of locating and applying a Filshie clip to a fallopian tube with an applicator:

Aspects of the process are depicted in greater detail in the following series of diagrams which appears in the manufacturer’s product information material:

The recommended location for the application of Filshie clips is over the muscular isthmic portion of the fallopian tube, between 1cm and 2cm lateral to the utero-tubal junction (or cornu). [5] The instructions contained in the manufacturer’s product information sheet include the following (emphasis added):

The product information sheet also contains the following instructions for intra-operative inspection of a closed clip (emphasis added):

Correct and incorrect compression and secure latching of the upper jaw or arm are illustrated in the following diagram:

When the clip is applied and fully closed, the curved upper jaw is straightened out and flattened. The closure of the two jaws applies pressure to the tube through the silastic linings, occludes it, and ultimately severs it. The expanding silastic linings apply constant pressure to the muscular portion of the tube, resulting in avascular necrosis [6] of the tube. With time, the silastic linings of each arm join together, resulting in complete transection [7] of the tube, involving destruction of approximately 4mm of the fallopian tube. This process takes some time to occur, and following it there is usually a defect between the proximal and distal ends of the fallopian tube, where the clip was affixed. At that stage, the clip becomes redundant, having achieved its objective of transecting and occluding the fallopian tube with ischaemic scar tissue. The manufacturer’s product information includes the following:

The resultant gap of about 4mm means that the clip is no longer attached to an anatomical structure and may migrate to virtually any area within the abdominal cavity. The possibility of such migration is enabled once the tube has been transected by the pressure of the clip and the process of avascular necrosis.

The following annotated diagrams, extracted from an article entitled “Failed sterilisation: evidence-based review and medico-legal ramifications” (“the 2004 Article”), [8] describe (on the left) the correct closure of a Filshie clip, and (on the right) incorrect under-closure (which the authors say is attributable to operator fault):

If a Filshie clip is left in an under-closed state – that is to say, it is not completely closed and locked – it might not completely occlude the fallopian tube, and it might dislodge from the tube, [9] and migrate to other structures – such as the broad ligament [10] – permitting fertilisation to occur.

It is not in doubt that pregnancy can occur after tubal ligation, without negligence. In a joint report, the expert witnesses – Professor O’Connor called by the patient, and Associate Professor Cooper called by the doctor – identified several possible reasons for a pregnancy occurring following a tubal ligation using Filshie clips, some of which did not involve operator error. [11] Dr Jeri gave evidence to the effect that where clips have been applied properly and fully locked across the width of the fallopian tubes, the failure rate after one year was 0.3%. [12]

The manufacturer’s product information brochure refers to a range of reported adverse events associated with Filshie clips, including uterine pregnancy (0.46%), ectopic pregnancy (0.16%), clip migration or expulsion (0.13%), and misapplication of clips to other tissues, namely ovarian ligaments, broad ligaments, omentum, bowel, tubal serosa, and cornual or broad ligaments (0.05%). This data does not enable negligent and non-negligent causes to be segregated. Nor does the manufacturer’s product information brochure identify the sample sizes from which it has been drawn.

However, most failed sterilisations involving Filshie clips, at least where the failure occurs within nine months of the procedure, are attributable to operator error. Professor Cooper’s second report, dated 13 June 2018, cited a 2007 article by Varma and Gupta entitled “Predicting negligence in female sterilization failure using time interval to sterilization failure: analysis of 131 cases” (“the 2007 Article”). [13] However, it annexed the 2004 Article, by the same authors. The 2004 Article was in evidence before the trial judge; due to a misunderstanding and confusion, the 2007 Article was not; it was admitted in evidence on the hearing of the appeal for reasons then given. [14]

In the 2004 Article, the authors referred to studies in several jurisdictions suggesting that Filshie clips were the most effective method of tubal ligation, with a failure rate of 0.84% (Canada), or between 0.11 and 1.93% (RCoG). One study showed the overall ten-year failure rate for Filshie clip sterilisations was 0.56% in 10,000 women, but this fell to 0.2% when cases caused by operator error were excluded. [15] This implies that the majority are attributable to operator error.

Failures not attributable to operator error were caused by:

The authors stated that, based on the data they obtained and analysed and a review of world literature, evidence for sterilisation failure revealed: [18]

In the 2007 Article, the authors analysed 131 failed sterilisations from the United Kingdom, Australia and a qualitative systematic review, of which 88 were attributed to negligence and 43 were non-negligent. They observed that the Filshie clip, most often failing due to non-occlusion or application to the wrong structure, was the predominant method in negligent failures (accounting for 71% of such cases); whereas the Pomeroy method, only failing by recanalisation and fistula, was the predominant method in non-negligent failures (accounting for 44% of such cases). Two non-negligent failures of Filshie clips were noted: one by fistula at fourteen months, and one by recanalisation at ten months. [19]

Relevantly, the authors concluded that sterilisation failure occurred significantly earlier in negligent than in non-negligent failure mechanisms (median failure intervals of 7.0 versus 12.0 months), and thus that knowing that sterilisation failure occurred ‘early’ increased the probability that the failure mechanism was likely to be negligent rather than non-negligent. Thus a ‘short’ interval to failure was suggestive of a negligent failure mechanism, with there being less certainty in the predictive value of longer time intervals on the mechanism of failure due to a paucity of cases. The authors wrote: [20]

They continued: [21]

Analysis demonstrated that 90% of failures occurring within nine months, and 77% of those occurring within twelve months, were negligent. [22] From the table of data it appears that the “tilting point” is at the nine-month mark: the majority of failures after nine months are non-negligent, although this conclusion is less reliable due to the paucity of cases.

While the 2004 article states that the possibility of tubal regeneration – either by fistula or recanalisation – had never been reported with Filshie or Hulka clips (as distinct from in sterilisation by electrocautery, fallope rings, or tubal excision), that must be modified in light of the 2007 Article. However, the Articles support the following relevant conclusions:

Unsurprisingly, the doctor did not claim to be able to recall the detail of the procedure she performed on 26 August 2014 – it was a procedure she had performed many times. She gave this evidence: [23]

She said that her usual practice when performing this procedure included “flipping” the tube after applying the clip, to ensure that it was completely closed over the tube. [24] She had not previously experienced a patient becoming pregnant subsequent to any tubal ligation, including by means of Filshie clips, performed by her. [25]

Whether on this occasion the doctor followed her usual practice was in issue. Ultimately his Honour concluded that she did not, and that conclusion is in issue in the appeal.

Intraoperative photographs of the application of the Filshie clips were taken by a nurse, and included photographs of the clips after they had been applied to the tubes. Photos 006, 007 and 009 depict the clip applied to the right tube, and photo 008 depicts one side of the clip applied to the left tube. [26]

Professor O’Connor and Professor Cooper agreed that the intraoperative photos demonstrated that the right clip had been correctly applied. They also agreed that the photographs did not conclusively show that complete closure of the left clip had been achieved, though neither suggested that they showed that closure had not been achieved. They agreed that they showed the left clip to have been placed more distally to the cornu than recommended – according to Professor O’Connor, about 2.9cm from the utero-tubal junction. Professor O’Connor said that the image appeared to show that additional vessels had been picked up with the fallopian tube, and that this had significance as the tube is the only structure that should be picked up, and the inclusion of extraneous tissue would reduce the likelihood of achieving correct occlusion, [27] and meant that the clip was not applied correctly, providing a possible explanation for the pregnancy, and some indication of incompetence in the application. Professor Cooper did not expressly disagree, but inferred from the apparent placement of the clips on the correct structures and the appearance that they were closed that there had been sufficient pressure for the left clip to be closed.

Whether the intraoperative photographs supported the doctor’s case that the left clip was properly closed, or (as his Honour ultimately found) accommodated the patient’s case that it was not, was and is a significant issue.

The caesarean section on 1 March 2016 was performed by Dr Jeri, coincidentally assisted by the doctor. After delivering the baby, Dr Jeri identified a clip on the right tube; he could not identify any clip remaining on the left tube. He performed a bilateral salpingectomy, [28] removing the whole of the right tube and about 80% of the left tube, including the clip on the right tube; the left clip, which he believed to be located “somewhere in the broad ligament”, was not retrieved. He said that there was no fistula.

The trial judge understood the effect of Dr Jeri’s evidence to be that prior to the salpingectomy, intraoperatively, both by visual examination and by palpation, he assessed the left tube as being intact along its length. [29] The appellant disputes that Dr Jeri’s evidence, properly analysed, was to that effect, and if it was, complains that it should not have been accepted when contrasted with the intraoperative photographs, the doctor’s usual practice, and the hysterosalpingogram next mentioned.

In the course of the litigation, on 20 August 2018, a hysterosalpingogram (HSG) (involving the injection of radio-opaque dye into the uterus) was performed and reported upon by Dr Stephenson, in order to ascertain whether there had been a recanalisation of the patient’s fallopian tube, or a fistula, and whether the tubes were patent or occluded. Dr Stephenson’s report contains a number of images taken in different planes, in which the left Filshie clip is visible. It was common ground between the experts that the test did not demonstrate any current fistula or recanalisation - although Professor Cooper said that the HSG did not preclude the existence of a fistula at an earlier time, which might not be evident due to scarring from the salpingectomy.

The HSG report contains eight images on which the left Filshie clip can be seen, in views taken in different planes. Originally, it was (incorrectly) interpreted as showing the clip in situ across and occluding the left fallopian tube, as the flow of dye appeared to end at the location of the tube. However, despite this earlier confusion, ultimately no party and neither expert contended that the clip was on the tube and obstructing contrast flow down the tube, and Professor Cooper conceded that the flow ceased at the point of the partial salpingectomy: [30]

Thus at HSG the clip was not on the tube, but it was in close proximity to the end of the post-salpingectomy remnant of the tube. The appellant contended that the HSG shows that the clip was closed. If so, that would weigh heavily, if not decisively, against the proposition that it had been incorrectly applied, as it would suggest that the clip remained closed and migrated having achieved its intended objective of transecting the tube.

In cross-examination, Dr Jeri said that he could not say with certainty whether the clip as it appeared on the HSG was open or closed; he said it may be closed, or it may be “partially open”. [31] Professor Cooper was of opinion that the clip depicted in the HSG imagery was closed (or, “exceedingly unlikely to be open”), [32] though he did not entirely exclude the possibility that it might have been incompletely closed. Professor O’Connor (whose opinion the trial judge accepted) held the view that the HSG images did not prove that the clip was closed and locked, and it remained a possibility that the clip was partially open, in the sense of not being completely locked - though it was “fairly clear that it’s not widely gaping open”, [33] but on this question would defer to a radiologist. [34] However, counsel for the doctor was not permitted by the trial judge to ask Dr Jones, a radiologist, to express an opinion in oral evidence, not covered by his report, as to whether the clip depicted in the HSG was open or closed.

Whether the clip so depicted was closed is a central issue: at the heart of the doctor’s case on appeal is a complaint that the finding that the left fallopian tube was intact along its length when examined by Dr Jeri prior to salpingectomy on 1 March 2016 is irreconcilable with the clip which had been applied to it having become dislodged from the intact tube without springing obviously open; whereas the appearance of the clip in the HSG images would be entirely consistent with its having performed its intended function by transecting the tube and then becoming dislodged and migrating elsewhere.

The trial judge found that the appellant placed the left clip in an incorrect location, [35] incorporated extraneous tissue in the bite of the clip, [36] failed to achieve full closure and locking of the clip, [37] and did not properly check that the left clip was completely closed and locked. [38] Those findings were inferences, drawn from the whole of the evidence. Although the twenty grounds of appeal relating to liability that were ultimately pressed attacked various steps in the trial judge’s reasoning, [39] ultimately the doctor must, if the adverse finding on liability is to be overturned, establish that the trial judge erred in concluding that, more probably than not, the patient’s pregnancy post tubal ligation was attributable to a departure by the doctor from proper professional practice in the application of the left Filshie clip, in particular by failing to achieve proper and complete closure of the clip.

In essence, the patient’s case was that operator error was to be inferred from the combination of:

In those circumstances, the patient contended, the more probable explanation of the whole of the evidence was that the clip was not fully closed, rather than that Dr Jeri was wrong. Against that, the contention advanced on behalf of the doctor was that a conclusion that more probably than not the clip was incorrectly applied could not be reached, having regard to:

The chief alternative explanation as to how the patient might have become pregnant which did not involve negligent application of the left clip was recanalisation, either direct or by fistula. No other theory was seriously advanced. The alternative explanations for the patient’s pregnancy were thus operator error (posited by the patient, and accepted by the trial judge), and recanalisation (posited by the doctor). There was no suggestion of non-negligent operator error: in other words, the case was conducted on the basis that if it were operator error, it was negligent operator error. Thus, in the way that the case was conducted, the patient’s pregnancy was caused either negligently by improper application of the left clip, resulting in incomplete occlusion due to failure to close it properly; or non-negligently by recanalisation.

Before turning to the overarching and dispositive question whether the trial judge erred in concluding that, more probably than not, the patient’s pregnancy post tubal ligation was attributable to a departure by the doctor from proper professional practice in the application of the left Filshie clip, in particular by failing to achieve proper and complete closure of the clip, it is convenient first to address two discrete categories of grounds of appeal: those which relate to the evaluation of witnesses, and those which relate to procedural rulings.

Ultimately this was a circumstantial case in which the critical conclusions depended on inferences drawn from the whole of the evidence, and not on the acceptance of any one witness in preference to another. That said, the trial judge made findings in relation to the credibility of a number of witnesses, which to some extent inform the weight to be given to their accounts, observations and opinions in the overall analysis of the circumstances to determine whether, more probably than not, they were explained by negligent operator error in the application of the left clip. The appellant challenges his Honour’s acceptance of Dr Jeri as reliable, his adverse evaluation of the credit of the doctor, and his preference for the opinions of Professor O’Connor over those of Professor Cooper.

Dr Jeri’s initial operation report of 1 March 2016 described in detail, in eighteen enumerated paragraphs, the conduct of the caesarean section, but referred to the salpingectomy much more briefly, merely stating after paragraph 18:

His original report did not refer at all to the location of the Filshie clips, or to the condition of the Fallopian tubes. However, he revised the report on 4 May 2016, and the revised report included, under “Findings”:

The description of the operation was amended to include the following:

There was still no reference to the condition of either Fallopian tube. However, the histological examination of the specimens removed at salpingectomy described the left tube specimen as “a defective segment of fallopian tube with attached fimbria 40x15x7mm”, and the right tube specimen as “a fallopian tube, 40mm x a diameter ranging from 10mm proximally to 15mm distally”. The diagnosis was, in respect of each tube, “Normal fallopian tube confirmed”. [41]

In his report dated 26 September 2018, prepared for the purpose of the proceedings, Dr Jeri answered questions asked in the instructing letter, which were not reproduced in his response. Interposing the relevant questions and answers, he stated:

In chief, Dr Jeri described his observations at the salpingectomy. He gave this evidence: [42]

Dr Jeri explained that while he performed a total salpingectomy on the right fallopian tube, he did only a partial salpingectomy on the left fallopian tube, including the fimbriae: [43] because of the scarring tissue, he removed only 80% of the left fallopian tube and fimbriae, removing the fallopian tube from the isthmus area down to the top of the fimbriae. [44]

With reference to his report of 26 September 2018, Dr Jeri was cross-examined as follows: [45]

With reference to Dr Jeri’s evidence, the trial judge said: [46]

Ultimately, the trial judge accepted Dr Jeri’s account. The appellant complains that his Honour erred in accepting Dr Jeri as reliable and preferring his evidence to that of the doctor (Ground 2), and in finding that Dr Jeri determined visually and by palpation along its length that the patient’s left fallopian tube was intact, when his evidence properly analysed did not demonstrate a determination as to the condition of the entire tube, so that the judge proceeded on an erroneous understanding of a material fact and used it as the foundation for findings on breach and causation (Ground 4).

The appellant submitted that Dr Jeri’s evidence was at best ambiguous as to the extent to which he ascertained that the left tube was intact along its length. I do not agree. From the passages set out above, it is clear that Dr Jeri said that he palpated, dissected and inspected the left Fallopian tube, which he observed was “absolutely patent and was of normal appearance”; that an area of damage from previous occlusion would have become apparent during that process but did not, and that the tube appeared normal (although conceding that without an HSG one could not be 100% certain) and not occluded; that if there had been a Filshie clip in that area, he would have seen it, but he could not locate the left Filshie clip, which was not on the tube, and which he assumed was somewhere “around the area of the broad ligament”; and that his answer that the left clip appeared intact was not based on intra-operative observation but on its appearance on the HSG, and meant only that it was not disintegrated, not that it was closed or locked, which he was unable to say.

It is true, as the appellant submits, that his Honour incorrectly took Dr Jeri’s answer to question 5 in his report of 26 September 2018, to the effect that the left Filshie clip appeared to be intact, to be a slip, and assumed that he had intended to refer to the left tube rather than the clip. Thus his Honour quoted the relevant paragraph of the report in the following terms: [47]

The matter “sic for fallopian tube” was his Honour’s insertion. Subsequently, his Honour recorded: [48]

However, as appears above, Question 5 in the request of 12 September 2018, to which Dr Jeri’s report of 26 September 2018 responded, was (emphasis added):

Moreover, the extracts from Dr Jeri’s evidence set out above confirm that answer 5 was intended to refer to the clip, although it was based not on intraoperative observation but on having seen the HSG. However, the trial judge’s misapprehension concerning answer 5 was not material, as Dr Jeri elsewhere – including in the same report – gave evidence to the effect of what his Honour incorrectly assumed answer 5 to state. Thus in the same report, it was stated in answer 7 that “the left fallopian tube was absolutely patent and was of normal appearance”. Dr Jeri confirmed in his oral evidence, set out above, that he was able to visualise and palpate the fallopian tube and detected no abnormality. Thus, although his Honour misapprehended that answer 5 was intended to refer to the tube and not the clip, the misapprehension was of no consequence.

Save for limited cross-examination as to whether he could be 100% sure, and that he had not mentioned the condition of the Fallopian tube in his report, Dr Jeri’s evidence about his inspection, palpation and dissection of the left tube was not challenged. Specifically, it was not put to him that the tube might have been divided, or occluded, yet he missed it. In those circumstances, his evidence was rightly understood to be to the effect that, on examination, palpation and dissection, the left fallopian tube appeared to be intact. Ground 4 therefore fails.

Ground 2, which impugns the acceptance of Dr Jeri in preference to the doctor and despite the HSG images, is ultimately intertwined with other grounds: Dr Jeri’s account forms part of the matrix of evidence from which the most probable hypothesis is derived. However, insofar as this ground deals discretely with Dr Jeri’s evidence, it is improbable, to a high degree, that Dr Jeri would have failed to observe that the tube had been divided or occluded if it had been, and it was not suggested to him that he might have. Dr Jeri described a quite extensive search for the clip on the left tube, including by palpation and dissection. The histology report – which describes the left tube specimen as a single specimen 40x10x7mm, and concludes that it was “normal” – tends to confirm that what was removed was not divided or occluded. This is reinforced by the HSG, which shows the remaining (post partial salpingectomy) proximal part of the left tube to be patent, even two years later, the dye stopping not at the clip but at the point where he had dissected the left tube two years earlier. [49]

As the appellant conceded, no direct challenge was made in cross-examination to Dr Jeri’s evidence concerning his findings on inspection, palpation and dissection of the left fallopian tube. Belatedly, in the course of submissions, his standing, expertise and impartiality were impugned; these attacks were justifiably rejected by his Honour, not least because they had not been raised in cross-examination. [50] The appellant sought to impugn Dr Jeri’s credit by reason that on 4 May 2016 – two months after it was made – he had revised his initial operation report, which did not refer to the location of the Filshie clips or the condition of the Fallopian tubes, to record that the left Filshie clip was located in the broad ligament, which was repeated in his report of 26 September 2018. This was said to be misleading because it was an assumption, not an observation. If this had any impact on Dr Jeri’s credit, it did not dictate that his evidence must be rejected. Not only was his assumption that the left clip was located in the broad ligament not argued to be an unreasonable one, but the doctor even claimed to have seen it there during the salpingectomy.

Taken as a whole, it was entirely open to accept Dr Jeri’s evidence as an accurate account of his intraoperative observations. The real issue about Dr Jeri’s evidence is its reconciliation with the HSG, considered later. [51]

The central question to which the doctor’s evidence related was the proposition that, in accordance with her usual practice, after applying the Filshie clips, she “flipped” the tube and checked that the clips were locked; [52] if so, that would weigh against a conclusion that she failed to use reasonable care. She also gave evidence that at the caesarean section, at which she was the assistant surgeon, she observed the left Filshie clip to be involved in adhesions on the left Fallopian tube.

As has been noted, the doctor unsurprisingly did not claim to be able to recall the detail of the procedure she performed on 26 August 2014. [53] As the trial judge said, the effect of her evidence in chief was that there was nothing outstanding or abnormal about this particular operation that caused her separately to remember details other than her usual procedure. His Honour, who saw and heard her give evidence, formed the impression that her evidence was essentially based on the limited content of her operation records, [54] and that her evidence in chief was to the effect that she assumed that she had carried out the procedure in accordance with her usual practice, which she proceeded to describe. [55] His Honour observed that her description of the operation appeared to be based on a general understanding of her usual approach to the procedure, and not an actual recollection of the events, details or observations, including that there was nothing that stood out as having occurred or having been observed which made it memorable. These findings, which are unremarkable, are not challenged.

Ultimately, the trial judge did not accept that, on this occasion, the doctor followed the usual practice she described. The appellant contends that his Honour erred in finding that the doctor did not follow her usual practice as described by her in applying Filshie clips, including “flipping” and checking them after application (Ground 14); and in making global adverse credibility findings based on peripheral issues, such as deficiencies in her note-taking and departures from her practice in respect of another matter (use of an information pamphlet) (Ground 20).

As to ground 14, the finding that on this occasion the doctor did not follow her usual practice does not depend solely on whether or not her evidence was regarded as credible, but on evaluation of the whole of the evidence, to which I return below.

As to ground 20, which impugns the trial judge’s assessment of her credit as a witness, the following matters at least, which cannot be passed off as “peripheral issues”, were adverse to her credibility, and also to her contention that she had on this occasion followed an invariable practice such as she described.

First, against the background of her evidence that there were no anatomical or other abnormalities that she would specifically recollect beyond reliance on her usual practice, in cross-examination she sought to justify the lateral placement of the left clip – in a place which did not accord with her usual (or recommended) practice – by the presence of a large varicose vein on the left fallopian tube. [56] Moreover, contrary to her evidence, the experts agreed that the images revealed no such varicose vein. [57]

Secondly, it emerged in cross-examination that she had only completed the sticker on the information pamphlet she said she had given the patient two years after the consultation, rather than contemporaneously. At the least, this showed that in that respect she had not abided by her usual practice. [58]

Thirdly, she at first claimed to have checked that she had achieved a proper closure by ensuring that a clicking sound was heard, and that she had heard an audible click when closing the clip, which she relied on as indication of complete clip closure. She later departed from this. The undisputed evidence was that the process in use at the relevant time did not produce an audible sound and that the relevant applicator did not produce a clicking sound (although a superseded version might have done so). [59] Professor Cooper agreed that reliance on an audible click would be an error. [60] Not only was the claim to have heard a click inconsistent with her lack of recollection of the detail of the operation; it was impossible.

Fourthly, the doctor said that at caesarean section the left clip was partially visible, involved in the adhesions on the broad ligament; she could not say whether the clip was open or closed. However, Dr Jeri – who as the operating surgeon was in a superior position to make an observation – said that despite palpation and dissection, he could not find the clip.

Taken together, these matters legitimately raised questions as to the doctor’s credibility, which was quintessentially a matter for the trial judge. Ground 20 fails.

His Honour explained, in comprehensive detail, [61] his reasons for preferring the opinions of Professor O’Connor to those of Professor Cooper, which were introduced and summarised as follows:

There followed a detailed exposition of each of the matters relied on, culminating in the conclusion:

There is no direct challenge to that conclusion which, being an evaluation of a witness who was cross-examined before him, was quintessentially a matter for the trial judge. However, the appellant contends that the trial judge erred in misstating the evidence to criticise and diminish the significance of Professor Cooper’s inability to reconcile the proposition that Dr Jeri had ascertained that the tube was intact with the objective evidence of the intraoperative photographs and the HSG, while failing to address the evidence of Professor O’Connor that the question of how a clip could come off an intact tube without springing open was “unanswerable” (Ground 10). The question of reconciliation of Dr Jeri’s evidence with the HSG, and Professor O’Connor’s evidence on the question of how a clip could come off an intact tube without springing open, do not depend exclusively on the evaluation of the exerts as witnesses, but form part of the overall analysis of the evidence and are addressed later. At this point, I address the ground insofar as it involved arguments that might undermine his Honour’s above-quoted conclusion that Professor O’Connor provided more measured and reliable expert guidance than Professor Cooper.

The appellant submitted that his Honour was unfairly critical of Professor Cooper in respect of a “newly emergent view that only arose during cross-examination”, in the following passage in the primary judgment: [62]

The appellant submitted that the distinction between a clip being “closed” and “locked” was in fact introduced by Professor O’Connor, in the first joint expert report. However, that does not detract from the point that Professor Cooper had not previously asserted that the clip was locked, and more significantly from the point that his readiness to state so categorically, in answer to the proposition that the intra-operative photo “doesn't reveal that the clip is locked though, does it?”, that “It’s locked” – not on the basis of the photo about which he was being asked, but based on a deduction that it would be obvious on the HSG images if it were open – reflected so firmly held an opinion as to betray a want of objectivity.

The appellant then submitted that the judge did not then set out further cross examination from which it was said to be clear that Professor Cooper disagreed with the proposition that the clip might not appear to be open, and thereby misstated the evidence. In that further cross-examination, Professor Cooper maintained that the clip was designed to be open and required pressure to close it, and that if it were surrounded by tissues and other organs in the pelvic cavity in contiguous proximity to the clip, they would not exert sufficient pressure to close it: [63]

It is not apparent to me how Professor Cooper’s opinion that on HSG the clip would have appeared clearly open if it were not locked is invalidated by the concession that the HSG does not show the left Filshie clip attached to and occluding the left tube, [64] but that (as he accepted) the clip could be surrounded by tissues and other organs in the pelvic cavity in contiguous proximity to the clip. To that extent, his Honour’s reasoning was inapt. However, that opinion, though not invalidated for the reasons his Honour supposed, still has to be weighed with other evidence, considered later, which was to the contrary effect.

The appellant also complains that the judge was critical of conclusions drawn by Professor Cooper from studies referred to by him as to the possibility of recanalisation, [65] while overlooking that both experts agreed that it was accepted that recanalisation in fact occurs, although it had not been reported in relation to Filshie clips, [66] and that the 2007 Article, referred to by Professor Cooper in his report of 13 June 2018, in fact reported failed sterilisation by way of fistula and also by way of recanalisation after Filshie clips. His Honour said:

The 2007 Article which reported two cases of failed sterilisation after Filshie clips, by way of fistula and recanalisation, was not in evidence before his Honour, though it was received on appeal, and though it was referred to in a footnote to Professor Cooper’s report, for another purpose, at trial no reference was made to this aspect of it. The absence from the above passage of specific reference to the agreement of the experts that it was “reasonably well accepted that the proximal and distal ends may re-join spontaneously by a process of recanalisation to allow transmission of oocytes and sperm” does not mean that his Honour overlooked it. To the contrary, his Honour specifically referred to it. [67] What was in contrast was Professor O’Connor’s acceptance that it probably sometimes happened, with Professor Cooper’s enthusiastic adoption of it as the explanation in the instant case. His Honour was explaining why the literature relied on by Professor Cooper did not persuade him that recanalisation was a likely explanation. His Honour did not overlook the agreement of the experts in their joint report that recanalisation was a theoretical possibility, but observed that the theory was ultimately and effectively abandoned by Professor Cooper in the course of the following evidence: [68]

Although in one respect I consider his Honour’s reasoning mistaken, that does not detract significantly from the conclusion, which was well-supported by the other matters to which his Honour referred, that Professor O’Connor’s opinions (most relevantly, ultimately, as to whether the clip was closed and locked), were more measured and qualified than those of Professor Cooper, whose objectivity was legitimately called into question by reason of his enthusiastic embracement of and adherence to a theory (recanalisation) which appeared to be an improbable explanation. It was reasonable for the trial judge to consider that Professor Cooper’s adamant certainty, contrasted with Professor O’Connor’s more measured approach, together with the other considerations to which his Honour referred, detracted from the reliability of the former’s evidence and opinions.

The appellant challenges two procedural or evidentiary rulings made by the trial judge in the course of the trial, namely that Dr Jeri would be permitted to express some expert opinions not contained in his statement, and that Dr Jones (radiologist) would not be permitted to give oral evidence not covered by his report, as to whether the HSG showed the clip to be open or closed.

Ground 3 complains that his Honour erred in accepting Dr Jeri’s opinion as an expert opinion over the objections of the appellant despite his not having been involved in the expert conclaves or the giving of joint expert evidence at the hearing and not serving a report containing the opinion he expressed in oral evidence.

When Dr Jeri was called, counsel for the doctor indicated that there was an objection to him giving evidence beyond his existing report. [69] It was clarified that the objection did not extend to factual evidence, but related to expert (opinion) evidence. [70] Thereafter, Dr Jeri proceeded to give evidence of his conduct of the caesarean section and salpingectomy, and the observations he made, without objection. Objection was taken to his giving opinion evidence as to the interpretation of the HSG. In permitting such evidence to be given, his Honour observed that any prejudice was more than able to be met by the period of time that would inevitably elapse between when the question was asked and answered and when the defendant would be in a position to call medical evidence. [71] His Honour’s discretionary judgment to permit that evidence to be given is unimpeachable.

As a result, Dr Jeri gave evidence to the effect that the HSG was indicative that the Filshie clip was not in place because part of the isthmic area could be seen, [72] and that the results of the HSG did not cause him to doubt the accuracy of the observations he had made at the time of the salpingectomy in 2016: [73]

That was the extent of his evidence in chief. It is inconceivable that he would not have been asked questions to that effect in cross-examination in any event based on the HSG, had he not been asked them in chief, to challenge his evidence as to his findings at the time of the salpingectomy. In any event, these aspects of Dr Jeri’s evidence do not feature in his Honour’s ultimate judgment.

As has been noted, in cross-examination, Dr Jeri said that he could not say with certainty whether the clip as it appeared on the HSG was open or closed; that it may be closed, or it may be “partially open”. [74] These opinions as to whether the clip was open or closed were elicited by the doctor’s counsel in cross-examination, in the course of seeking unsuccessfully to obtain a concession that the clip appeared to be closed – when no opinion on that topic had been adduced in chief. The doctor cannot complain that Dr Jeri gave adverse opinions in response to questions asked in cross-examination by her own counsel.

This ground fails.

As has been noted, Professor O’Connor indicated that on the question of whether, on the HSG imagery, the clip appeared to be open, the opinion of a radiologist would be preferable. Ground 12 complains that the trial judge erred in refusing to permit oral evidence in chief to be adduced from Dr Jones regarding the appearance of the clip in the HSG imagery.

When Dr Jones, the radiologist who had reported on the HSG at the request of the doctor, was called, the doctor’s counsel endeavoured to ask him in chief: [75]

Objection was taken to that question, including on the ground that his report did not cover that matter. The trial judge asked whether there were “exceptional circumstances” that justified leave being given to adduce the evidence for the purposes of UCPR r 31.5, and counsel referred to it being a matter that had been addressed by other experts, including the plaintiff’s expert, and that it was a material matter. [76] Counsel for the patient responded, inter alia, that the patient had not retained an expert radiologist because the doctor was not adducing expert radiological opinion. His Honour refused leave. [77] His Honour’s discretionary decision to refuse leave is unimpeachable: the matters referred to by the doctor’s counsel hardly amounted to “exceptional circumstances”, whereas that identified by the patient’s counsel was one involving real prejudice if leave were granted.

Although it was contended that the refusal of leave for the appellant to adduce evidence on this question from Dr Jones stood in contrast to the judge having permitted Dr Jeri to opine, having regard to the HSG, whether the clip was open or closed, that was in the context that, as his Honour observed, [78] any prejudice was more than able to be met by the period of time that would inevitably elapse between when the question was asked and answered and when the defendant would be in a position to call medical evidence. That was not the case when Dr Jones was called. Moreover, as has been explained, the relevant evidence was elicited from Dr Jeri not in chief, but in cross-examination.

No reference was made in argument before his Honour to the significance now attributed to Professor O’Connor’s deference to the opinion of a radiologist on this question. That was no doubt because Dr Jones was called before Professor O’Connor. No application to recall Dr Jones was made after Professor O’Connor indicated, in substance, that the question was really one for a radiologist.

This ground fails.

The remaining grounds of appeal were to the effect that his Honour erred:

Notwithstanding the number and detail of these grounds, they cannot avail the appellant unless they vitiate the trial judge’s conclusion that, more probably than not, the patient’s pregnancy post tubal ligation was attributable to a departure by the doctor from proper professional practice in the application of the left Filshie clip, in particular by failing to achieve proper and complete closure of the clip. Essentially, these grounds attack various aspects of his Honour’s reasoning underpinning that conclusion. As has been noted, that conclusion was an inference drawn from the circumstances. As the competing cases of the parties, earlier summarised, [79] indicate, the significant integers were:

In a circumstantial case such as the present, it is the ultimate question – not the various intermediate facts – which must be proved according to the balance of probabilities. A set of circumstances may be explained, more probably than not, by an hypothesis which includes matters which, standing alone, seem improbable. Thus it is essential not to be distracted by irrelevant questions on subsidiary issues, such as whether on the HSG (or on the intraoperative photos) it was more probable than not that the clip was open; or whether on Dr Jeri’s evidence it was more probable than not that he found the tube to be intact. That is because resolution of the question whether, more probably than not, the clip was not properly closed, depends on the mutually supporting effect of the whole of the circumstances, and not on one or more of them in isolation.

That is subject to this exception. An hypothesis otherwise capable of explaining the circumstances may be excluded if it appears to be inconsistent with an irrefutable fact. Thus the hypothesis which his Honour ultimately accepted – that the doctor failed to achieve complete closure of the left clip – would have had to be excluded if the intra-operative photographs conclusively demonstrated that complete closure was achieved, or if the HSG conclusively showed that the clip remained fully closed. However, if the evidence fell short of being conclusive of those issues, the hypothesis remains an available one, to be considered and weighed with the alternatives.

In their joint report, Professor O’Connor and Professor Cooper identified the following possible reasons for a pregnancy following a tubal ligation using Filshie clips:

The most common reason for a clip migrating from the tube is if it is not properly locked, and “then it just falls off … [leaving] patent the fallopian tube”. [80] If the clips are locked properly, they do not open spontaneously; [81] but if not locked properly, the clip can dislodge, which is the main reason for the procedure not being effective. [82] The academic literature to which reference has been made is indicative that the most likely cause of failure of a tubal ligation with Filshie clips – particularly if pregnancy occurs in the first nine months after the procedure – is operator error in application of the clip; and that under-closure is a known cause for failure, that is generally attributable to operator error.

As has already been noted, in the present case, the foremost alternative explanation as to how the patient might have become pregnant which did not involve negligent application of the left clip was recanalisation, either direct or by fistula. While this was proposed as a possible explanation by Professor Cooper in his first report, it originated from his misinterpretation of the post-salpingectomy histology as reporting a complete transection of the left tube by the clip, on which basis he concluded that this was highly suggestive of tubal lumen regeneration. In his oral evidence, extracted earlier, [83] he adhered to fistula as the explanation.

As has also been explained, it was the doctor’s evidence that, in accordance with her usual practice, after applying the Filshie clips, she “flipped” and checked that the clips were locked; and that at the caesarean section, at which she was the assistant surgeon, she observed the left Filshie clip to be involved in adhesions on the left Fallopian tube. In the absence of an actual recollection of how she performed the operation and what she saw and did, evidence of the doctor’s usual practice provides some evidence of what she might have done on the occasion in question. However, it has to be weighed with other evidence from which inferences are available. Moreover, as has been explained, there were several matters which significantly detract from the proposition that on this occasion the doctor followed an invariable practice such as she described. Of course this does not of itself mean that more probably than not the clip was not properly applied; but it leaves open, as one of the available explanations of the circumstances, the possibility that she did not follow that practice.

Professor O’Connor’s opinion was that photo 008 showed the clip on the left tube to be located about 2.9cm from the utero-tubal junction (or cornu), on the mid portion of the left tube rather than on the isthmus, and appeared to show that some extraneous tissue, including the leash of left ovarian blood vessels, was caught in the bite of the left clip in addition to the tube, which was an incorrect application as the bulky inclusion of adjacent anatomical structures would reduce the likelihood of achieving correct occlusion. [84] Moreover, this inclusion of extraneous tissue would provide an explanation as to why the clip could fall off the tube, and some indication of incompetence in the application. This evidence was not contradicted.

In their first joint report, Professor O’Connor and Professor Cooper agreed that the intraoperative photos demonstrated that the right clip had been correctly applied. They also agreed that the photographs did not show with certainty that either clip was locked in place, although neither suggested that they demonstrated that they were not locked. Professor O’Connor said that the image appeared to show that additional vessels had been picked up with the fallopian tube, the significance of this being that the tube is the only structure that should be picked up, and the inclusion of additional vessels meant that the clip was not applied correctly and provided a possible explanation for the pregnancy. Professor Cooper did not expressly disagree, but inferred – from the apparent placement of both clips on the correct structures and the appearance that they were closed – that sufficient pressure had been applied for the clip to be closed.

In oral evidence, Professor O’Connor said, regarding the clip, that: [85]

Professor Cooper agreed that the left clip was “lateral to where the preferred position would be, yes”, [86] but speculated “Look, I think there’s reasons as to why it could be placed in a different position”. [87] He agreed that the photographs did not show both sides of the left clip, and did not show the jaws to be locked, but said “I think you can infer that”. [88] He explained that he inferred – from the “slightly swollen” appearance of the muscular portion of one side of the tube adjacent to the clip, which he also described as “a degree of expansion of the tube lateral to the clip”, “insinuating to me that that clip was closed” – that there was “some pressure at that point [which] reasonably convinced me that that clip was closed”. [89] However, he agreed that he could not tell from the intraoperative image whether or not there had been under-closure of the clip; and that he could not exclude the possibility that the clip could have been under-closed. [90]

The trial judge concluded: [91]

Self-evidently, that paragraph refers back to the analysis contained at primary judgment [495] to [561], discussed earlier, [92] in which his Honour sets out in detail his reasons for doubting Professor Cooper’s objectivity. The appellant’s criticism that his Honour did not identify any basis for the conclusion expressed as to his reliability is unfair.

It was common ground between Professor O’Connor and Professor Cooper that the intraoperative photographs did not conclusively show that complete closure of the left clip had been achieved, though neither suggested that they showed that closure had not been achieved. Professor Cooper’s opinion that the clip was locked depended on inference from the associated observation of the “dilated” appearance of an area on the proximal side of the clip which appeared to have been compressed from which he inferred that the clip was closed. [93] However, he agreed that this was a qualified and not a definitive view. [94] Moreover, it is not apparent, and the evidence did not establish, that the “slightly swollen” appearance resulting from compression that he described could not be associated with an under-closed clip (perhaps also associated with the incorporation of other tissue) as well as with a fully closed clip. Under-closure applies pressure to and partially occludes the tube, at least so long as the clip remains in place, so logically it too might be expected to produce some dilation proximal to the location of the clip. Indeed, Professor Cooper acknowledged this (emphasis added): [95]

Thus he agreed that he could not tell from the intra-operative photographic image of the left Filshie clip whether or not there had been under-closure of that clip, and that it was possible that this was a case of under-closure. [96] As the above quoted passage also illustrates, Professor Cooper’s confidence that the clip was closed was also clearly influenced by his interpretation of its appearance on the HSG, [97] which is discussed later. [98]

It was also common ground that the intraoperative photos showed that the left clip was placed more distally to the cornu than recommended, and Professor O’Connor’s evidence that the image appeared to show that additional vessels had been picked up with the fallopian tube, meaning that the clip had not been correctly applied and providing a possible explanation for the pregnancy, was unchallenged and uncontradicted.

Thus, the intraoperative photos did not exclude the possibility that the left clip was not properly and completely closed. Moreover, in demonstrating that the clip was placed more distally to the cornu than recommended, and included extraneous tissue, they provided a potential explanation for why complete closure and occlusion might not have been achieved, increasing the plausibility of the hypothesis that the patient’s pregnancy was attributable to incomplete closure of the left clip.

As has been explained, [99] Dr Jeri’s evidence was to the effect that on inspection, palpation, and dissection, the left Fallopian tube was absolutely patent, of normal appearance, and intact; and he could not locate the left Filshie clip, which was not on the tube, but which he assumed was somewhere around the area of the broad ligament. He also said that there was no fistula. That evidence, which was not challenged, both told against the clip having transected the tube, and against the alternative theory of fistula. To some extent, Dr Jeri’s evidence was supported by the absence of any abnormal findings in the post-operative histology. [100]

The appellant contends that the trial judge erred in accepting Dr Jeri’s evidence in preference to that of the doctor. However, except on the minor issue of whether the Filshie clip was observable at salpingectomy, there was no direct conflict between them. The doctor’s evidence which was not accepted related to conduct of the tubal ligation in accordance with her usual practice, whereas Dr Jeri’s which was accepted related to his observations 18 months later at salpingectomy. In circumstances where his evidence was not challenged, it is hardly surprising that it was accepted. The appellant’s argument appears to be that it was inconsistent with irrefutable facts, in particular the appearance of the clip on the HSP; however, as explained below, that is not the case.

Dr Jeri’s evidence as to what he observed at salpingectomy is not conclusive of the ultimate question, and it must be weighed with the other evidence, but it weighs considerably in favour of the theory that the clip was not properly applied, because it is indicative that the tube had not been transected by the clip, and that there was no fistula.

The appellant contends that the HSG shows that the clip was closed. If so, that would weigh heavily, if not decisively, against the proposition that it had been incorrectly applied: no theory, let alone a plausible one, has been suggested as to how it might have become closed by the time of the HSG if it were not originally closed.

One important element of the contention for the doctor that the clip as it appeared on the HSG was closed and locked was the proposition that if it were not locked, it would have sprung wide open and that this would be apparent on the HSG. It was asserted that a Filshie clip is designed so that if it is not properly locked by the self-calibrating applicator, it will spring back open.

No statement to that effect appears in the product information. Notably, the diagrams depicting incorrect or under-closure (reproduced at [12] and [15] above) are suggestive that an under-closed clip does not necessarily “spring” open.

The appellant referred to two evidentiary passages as supporting the proposition. The first was the doctor’s evidence: [101]

The second was the following evidence of Professor Cooper: [102]

It is clear enough that before the clip is closed the expansive effect of the silastic insert exerts pressure that keeps the arms of the clip in the open position, so that it requires the application of force to close the clip. However, whether it has that effect after partial closure, involving some but incomplete straightening and flattening of the upper arm, is not clear. Asked how, assuming that the clip was no longer on the tube at salpingectomy, that could have occurred other than by migration “in the recognised way” (a reference to migration following the clip transecting the tube) or by the clip opening and coming off the tube, Professor Cooper said that if the clip was open, an open clip would have been seen on the HSG. While he said that if the clip was open it would be obvious on the HSG images, [103] he appeared unsure as to whether the clip’s upper arm could or would spring (or “bounce”) back open after the applicator had bent it during deployment, stating (emphasis added) “… if it can bounce sort of open then I would expect it to be obvious on the … HSG. But the HSG appears to reveal [it] to be closed”. [104] And when asked how the clip could become detached from an intact tube without springing open or cutting through the tube, Professor Cooper offered that it might have been “loose but not widely open”. [105]

Professor O’Connor explained that if the clip were partially but incompletely closed, it would not spring open completely, because the architecture had been changed by the flattening of the arm: [106]

Professor O’Connor later gave this evidence: [107]

In response to that, Professor Cooper said (emphasis added): [108]

In my view, the evidence did not establish that, if it had been applied but incompletely closed, the clip would necessarily spring wide open to the condition in which it was before application.

In a joint expert report of 28 March 2019, the experts agreed that in the HSG imagery, the dye was filling “to the level of the clip on the left hand side”. (At that time, the belief that the HSG depicted the clip affixed to the tube was still current). They gave the following opinions:

In respect of this, Professor O’Connor gave the following oral evidence: [109]

He also said: [110]

Both experts agreed that it would be difficult to differentiate on the HSG images between a partially but incompletely closed clip, and a completely closed clip. [111]

As has been noted, Professor Cooper, having agreed that he could not tell from the intraoperative image whether or not there had been under-closure of the clip, conceded that though he would expect the HSG to appear differently if the clip was not applied correctly and resulted in under-closure, it was nonetheless a possibility. [112]

As also noted above, [113] Professor Cooper said that if the clip was open it would be much more obvious on the HSG images, [114] and that if “it can bounce sort of open” then he would expect it to be obvious on the HSG, on which it appeared to be closed. [115] As has also already been recorded, [116] he also said that the clip appeared to him to be closed, contending that if it had not been properly closed “it would presumably spring back open again”, but acknowledged the possibility that it might be distorted or “sort of half open, if you will, or not quite closed” which would not be picked up on the HSG. [117]

Professor O’Connor responded, “I don’t think I have much more to add, your Honour”, [118] but offered this qualification: [119]

Professor Cooper maintained that it was “exceedingly unlikely to be open”. [120]

Later, as has been recorded, [121] Professor O’Connor confirmed that in his view it remained a possibility that it could be “partially” open and “not completely locked in” – though it was clearly “not widely gaping open”, [122] to which Professor Cooper replied that he agreed “in large measure in the sense that it doesn’t appear to be widely open”, acknowledging that there might well be “different stages of incomplete closure depending upon the force applied to upper arm”, resulting in partial compression of the silastic material. [123]

Undoubtedly the effect of the evidence was that the HSG did not show the left clip to be fully open. However, while the experts differed in degree, and Professor Cooper certainly leant in favour of the view that it was closed, both ultimately accepted that there was a possibility that it was incompletely closed. Professor Cooper’s confidence that the clip was closed was largely based on the unsubstantiated presumption that the clip would “spring back open”. In any event, he ultimately acknowledged the possibility that it might be partially but not completely closed, in a way which would not be apparent on the HSG.

Thus the HSG imagery did not exclude the possibility that the clip was incompletely closed. The trial judge rightly concluded that the HSG images did not permit one to say that the clip was either open or fully closed. [124] That is not to cast any burden of proof on the doctor to show that it was closed, but to establish that the HSG imagery was not conclusive, and had to be weighed and reconciled with all the other evidence in resolving whether, more probably than not, the clip had been inadequately closed.

At the core of the appellant’s case was the argument that no mechanism had been identified whereby the clip would not be obviously open if it had not been correctly applied to occlude the tube. It was submitted that the judge failed to identify how the clip could have become separated from the tube without springing open, or how it could move, migrate or dislodge while remaining closed as was said to be demonstrated by the HSG, particularly in light of evidence that if closed but not locked the clip should spring back open, and Professor O’Connor’s supposed inability to provide an explanation for how it could detach without springing open or cutting through the tube.

However, the force of that argument is much diminished once it is appreciated that, as has been discussed, the evidence did not establish that the clip would necessarily “spring back open” after it had been partially closed, nor exclude the possibility that the HSG depicted a clip that was partially but incompletely closed. Moreover, the mechanism by which it could be dislodged from the tube was not unexplained in the evidence.

The appellant relied on the following passage in Professor O’Connor’s evidence for the proposition that he was unable to explain how the clip could detach from the tube without springing open unless it had transected the tube as intended: [125]

The answer was provided by Professor Cooper, as follows: [126]

I do not understand Professor O’Connor to be saying that the appearance of the clip on the HSG and its detachment from the tube was inexplicable or irreconcilable, but rather that he was unable to prove how it had happened. In his report of 20 June 2017, Professor O’Connor had explained (emphasis added):

Dr Jeri provided a similar explanation: [127]

In my view it is tolerably clear that, although they may have differed as to the likelihood of it in this case, all three of Professor O’Connor, Professor Cooper and Dr Jeri were saying that it was possible for a Filshie clip that was not completely closed to dislodge from the Fallopian tube, leaving it unoccluded. As already explained, this would not necessarily involve it springing fully open.

Contrary to the appellant’s submissions, the trial judge did identify mechanisms by which this might occur: [128]

His Honour then proceeded to summarise those explanations, which included disturbance of the clip when it was being “flipped” for inspection, or by intra-abdominal organ pressure in the context of a re-establishment of the normal anatomical apposition of organs and structures following removal of the insufflating gas that had been introduced into the abdominal cavity in order to separate those structures for the purpose of identifying and isolating them in order to perform the tubal ligation procedure; or that the doctor may have believed she was closing the left Filshie clip in her chosen position on the left fallopian tube, when in fact she may have incompletely closed it whilst including extraneous non-fallopian tube tissue in addition to the tube, in an apparently but not actually closed state, with the result that the left clip subsequently dislodged from that tissue.

As has been mentioned, it is the ultimate question – not the various intermediate facts – which must be resolved according to the balance of probabilities. In this case, resolution of whether, more probably than not, the failure of the tubal ligation to prevent the patient’s pregnancy nine months after the procedure was attributable to operator error in application of the left clip involves evaluating potential explanations each of which, standing alone, might appear unlikely: recanalisation (directly or by fistula) of the left tube on the one hand, and under-closure of the left clip on the other. However the evidence, including the academic literature, is indicative that, all else being equal, such a failure is most likely to be attributable to operator error. In his second report, Professor Cooper acknowledged that the short interval in this case between application and failure was suggestive though not indicative of a negligent event.

In the absence of an actual recollection of how she performed the operation and what she saw and did, the doctor’s evidence that, in accordance with her usual practice, she “flipped” and checked that the clips were locked, provides some evidence of what she might have done on the occasion in question. However, as has been explained, [129] there were several matters which cast doubt on whether on this occasion the doctor followed an invariable practice such as she described. Her evidence has to be weighed with other evidence from which inferences are available.

The positioning of the clip outside the recommended distance from the cornu, and in particular its inclusion of extraneous tissue in addition to the fallopian tube, provides additional reason for supposing that the pregnancy was attributable to incomplete closure. Contrary to the appellant’s submission, the evidence did not establish that the clip was applied to an anatomically appropriate portion of the tube and was adequately closed: Professor O’Connor’s evidence, as has been summarised above, was clearly to the contrary; and Professor Cooper accepted that it was possible that the clip was not completely closed.

The alternative explanation of recanalisation or fistula, in the context of this case, also necessarily involves the clip having become detached from the tube not because of incomplete closure, but because it transected the tube as intended. In their first joint report, the experts agreed that spontaneous recanalisation by the proximal and distal ends of the tube re-joining had not been reported with Filshie clips, and was less common than re-joining through a tubal fistula, but that it was “reasonably well accepted” that, notwithstanding the appropriate placement of clips, the proximal and distal ends of the fallopian tube may ultimately recanalise. However, recanalisation or fistula typically requires some twelve months or more. As has been noted, the 2007 Article notes two non-negligent failures of Filshie clips – a fistula at fourteen months, and a recanalisation at ten months. Thus there is no reported precedent for recanalisation or fistula accounting for a pregnancy at nine months after the procedure.

Moreover, the HSG was undertaken to ascertain whether there was a fistula, and whether the tubes were patent or occluded. It was common ground between the experts that the test did not demonstrate any current fistula or recanalisation, and although Professor Cooper, whose view was the pregnancy must have been caused by a fistula, said that the HSG did not preclude the earlier existence of a fistula which might not be evident due to scarring from the salpingectomy, Dr Jeri, who performed the salpingectomy, said that there was no fistula then. Further, Dr Jeri’s evidence to the effect that the left tube was found to be intact on 1 March 2016 told against the proposition, necessary to the alternative theory, that the clip had become detached from the tube not because of incomplete closure, but because it transected the tube as intended.

In all those circumstances, recanalisation or fistula appears a most improbable explanation.

Thus the mere fact of failure, occurring as early as it did, was suggestive of operator error in application of the clip. That hypothesis was reinforced by the circumstance that the left clip was applied at a sub-optimal point on the tube and incorporated extraneous tissue, increasing the chances of an incomplete closure. The appearance of the clip in the HSG images did not exclude that hypothesis, because the clip could have been incompletely closed. The alternative explanation of recanalisation or fistula is rendered highly improbable by the relatively short time between operation and pregnancy, and by the evidence of Dr Jeri and the HSG, all of which militated against a fistula circumventing an otherwise successful transection. More probably than not, the patient’s falling pregnant nine months after the tubal ligation is in those circumstances to be explained by operator error on the part of the doctor in the application of the left Filshie clip.

Contrary to the appellant’s submissions, this approach – and that adopted by the trial judge – involves no “reversal of the onus”. The appellant contended that the judge erred in “reversing the onus of proof” by finding that the doctor could not demonstrate that the left Filshie clip was completely closed and locked over the tube, in circumstances where the evidence did not demonstrate that it was not (Ground 1(a)); by finding that the appellant could not demonstrate that the left clip as depicted on HSG was closed, when the evidence could not demonstrate that it was open (Ground 1(b)); and by not accepting the evidence of Professor Cooper and the appellant and finding that that meant the opposite of their evidence had been proven and that the evidence of Dr Jeri and Professor O’Connor could be accepted without further analysis (Ground 1(c)).

These complaints must be seen against the background that, but for the appearance of the clip on the HSG suggesting that it was closed, and the doctor’s evidence as to her usual practice, the evidence very strongly pointed to a conclusion that the clip was not properly applied. The circumstances were such as to call for an explanation. It is a misconception that the judge’s finding impugned by Ground 1(a), that there was no satisfactory evidence to show that the appellant had achieved full closure and locking of the left clip because of the absence of reliable photographic evidence, [130] involved any reversal of the onus of proof: his Honour did not reason that because the clip was not shown to be closed, it was therefore open. The significance of the photographic evidence was that the inference that the doctor had failed to achieve full closure was not excluded by the intraoperative photographs and remained an available one, to be weighed with the alternative hypotheses in the light of the whole of the circumstances.

Similarly, the finding impugned by Ground 1(b), that the left clip as depicted on the HSG was not shown to be closed, involved no reversal of the onus: his Honour did not reason that it was therefore open. Again, it simply meant that the hypothesis (that the doctor had failed to achieve complete closure) remained available, to be weighed with the alternatives.

Ground 1(c) is likewise misconceived. His Honour did not reason that it followed, from the rejection of the evidence of the doctor and of Professor Cooper, that the opposite had been proven. The conclusion that on this occasion the doctor did not follow her usual practice in applying and checking the clips does not depend solely on whether or not her evidence is regarded as credible, but on evaluation of the whole of the evidence. His Honour affirmatively preferred the evidence of Dr Jeri and of Professor O’Connor, and concluded that the circumstances were most probably explained by the doctor having failed to achieve complete closure of the left clip, as appears from the following passage in his Honour’s reasons, which also refutes the complaint that his Honour did not have regard or proper regard to the HSG imagery: [131]

For the foregoing reasons, the trial judge was right to conclude that on the balance of probabilities, the patient’s pregnancy post tubal ligation was attributable to negligent operator error on the part of the doctor in the application of the left Filshie clip, in particular by one or more of placing the clip too distally and including extraneous tissue in the bite of the clip (as appeared from the intraoperative photograph), failing to achieve full closure and locking of the clip (an inference from the circumstances), and/or failing properly to check that the left clip was completely closed and locked (as such a check should have revealed that it was not). The contention that the judge erred in failing, in a circumstantial case, to have regard to the fact that the appellant had not previously had a failed sterilisation following application of a Filshie clip, is without substance: it is no more relevant than would be the fact that, in a motor vehicle case, the defendant driver had not previously had an accident.

In light of the conclusion which I have reached, it is unnecessary to consider the Notice of Contention by which the respondent seeks to uphold his Honour’s conclusion on alternative grounds.

The trial judge assessed damages in the total sum of $408,700, calculated as follows:

Grounds of appeal 23 and 24 complain that the trial judge erred in interpreting Civil Liability Act 2002 (NSW) (“CLA”), s 71, as not precluding the patient’s claim for loss of earnings (Ground 23); and in assessing her claim for non-economic loss at 38% of a most extreme case, and awarding damages for economic loss “in breach of s 71”, so that her total damages were wholly disproportionate to the alleged harm caused (Ground 24).

In the appeal, the doctor does not contest the calculation of the economic loss components but submits, as was submitted at trial, that it is excluded by CLA s 71(1)(b). Section s 71 is located in Part 11 of the Act. Section 70 relevantly provides that Part 11 “applies to any claim for damages in civil proceedings for the birth of a child”. Section 71 provides as follows:

The patient’s claim for past financial loss was founded not on her giving up work (or refusing promotion) to look after the child, but on her ability to work being constrained by on-going mental health consequences originating from the discovery that, despite the tubal ligation, she had conceived a child. It covered:

In respect of lost future earning capacity consequent upon her compromised mental health, the patient claimed a buffer of $100,000. The judge allowed $80,000.

On behalf of the doctor it was submitted that:

Essentially, the contention advanced on behalf of the doctor is that, on a literal construction of s 71(1)(b), the width of the term “any loss of earnings” and the temporal connector “while” means that any loss of earnings the patient suffered during the time that she was involved in rearing or maintaining the child is irrecoverable.

Section 71 was enacted in response to the decision of the High Court in Cattanach v Melchior, [133] with the intention of altering the common law position as established by that case, in which Mrs Melchior underwent a tubal ligation after she and her husband decided that they wanted no more children. As a result of negligence in the performance of the procedure, she gave birth to a healthy son. At trial (1) Mrs Melchior recovered damages for pain, suffering and loss of amenities of life associated with the pregnancy and childbirth; loss of some part-time earnings; loss of earning capacity because of a thrombosis associated with the pregnancy; and medical and pharmaceutical expenses; (2) Mr Melchior recovered damages for loss of consortium during his wife’s pregnancy; and (3) Mr and Mrs Melchior jointly recovered the costs of raising and maintaining the child to age 18. Only the third category of damages was in issue in the High Court, which by majority upheld the award of damages for the costs of raising the child, essentially on the basis that once it was established that the birth of the child had been caused by the defendant’s negligence, there was no proper basis for denying the parents damages in respect of the foreseeable expenses arising from that event. [134]

The legislative response encapsulated in CLA s 71 was informed by the “strong moral objection to such damages because they classify the birth and existence of a child as an ‘injury’ to the child’s parents”. [135] The intention was to deny a claimant damages for the costs of raising a healthy child and “damages for any loss of earnings suffered by the parents while rearing or maintaining the child”, “for example, if a parent gives up work or refuses promotion to look after the child”. [136]

Section 71 is a limited intrusion on the common law and (contrary to the appellant’s submission) does not involve denying a mother damages for the pregnancy and birth of the child, but only for the costs of raising the child, and lost earnings “while the claimant rears or maintains the child”. The Second Reading speech explicitly acknowledged that where a pregnancy is a result of negligence, the mother would still be entitled to recover damages for the pregnancy and the birth, though not for the cost of raising the child. [137]

The first (pre-natal) component of the patient’s claim for past economic loss is on no view caught by s 71(1)(b), which refers to the period “while the claimant rears or maintains the child”, and does not apply to the period of pregnancy before the birth of the child.

Nor is the second (four week post-natal) component: during that period, the patient was unable to breast feed, and her mother-in-law cared exclusively for the child; as she was not on any view rearing or maintaining the child during this period, it would not be caught by s 71(1)(b).

As to the third component of the claim for past economic loss, and the claim for future loss, the policy underlying s 71 is the moral judgment that parents of a healthy child should not be entitled to recover damages for raising, rearing and maintaining that child. That policy does not require that recovery be denied in respect of the loss of earnings which, though coincident with the raising of a child, is attributable not to the need or choice to raise the child, but to another cause, such as a physical or psychological injury (albeit one that is connected with the birth of the child). The phrase “while the claimant rears or maintains the child” is intended to capture loss of earnings arising from the need or choice of a parent to rear the child and its consequent impact on the parent’s income, as distinct from lost earnings attributable to a disability arising from the pregnancy or birth. It follows that, in proceedings involving a claim for the birth of a child, CLA s 71(1) does not preclude the award of damages for economic loss for loss of earnings by the claimant attributable to psychiatric injury associated with the birth of the child, as distinct from a need or choice to rear or maintain the child.

The patient’s case was that she had returned to work approximately five months after the child was born for one day a week, but she was not mentally capable of working more than one day a week, as she would have had she been mentally capable. Thus her inability to work to the full extent of her capacity arose not from the fact that she was rearing or maintaining the child, but because she was incapable of working any more than she did due to her psychiatric condition. [138] No ground of appeal challenges the judge’s finding that the patient had developed a major depressive disorder that had worsened and had been pervasive since the child was born, and which to fluctuating degrees would be with her for years. [139] Nor does any ground of appeal challenge the judge’s finding that she has in fact incurred loss of earnings and loss of earning capacity due to her psychiatric illness that followed the conception, pregnancy and birth of the child. [140] Importantly, in assessing damages, his Honour took into account as a vicissitude the possibility that the patient’s earning capacity might in any event have been impacted by rearing or maintaining the child.

This ground therefore fails.

This ground contends that the trial judge erred in assessing the respondent’s claim for non-economic loss at 38% of a most extreme case “and awarding her economic loss in breach of s 71 … so that her total damages were wholly disproportionate to the alleged harm suffered”. Insofar as this ground impugns the award for economic loss, said to be “in breach of s 71”, that has been addressed under the preceding ground. That then leaves the question of the assessment of non-economic loss at 38% of a most extreme case.

The trial judge accepted the patient’s written submissions as aptly summarising her position, including her “horrible” experience of the birth; that after the birth she became very withdrawn from her husband, her social circle and life generally; that she had been unable to recommence exercise and her libido had been affected by antidepressant medicine, and her weight increased; that she continued to be traumatised and suffer from symptoms consistent with ongoing psychiatric disorder including panic attacks; that her self-esteem was adversely affected, and she still cried at times, though this had been ameliorated by the antidepressant medication which she continued to take; that she was very angry about what had occurred, and though hopeful that she would get better there had been no improvement; and that the psychiatric experts agreed that she suffered from a major depressive disorder and that there was ongoing impairment. [141] His Honour characterised these matters as describing “a significant and life-changing event that has had and will continue to have an adverse impact on the plaintiff’s autonomy and her plans for her life, and on her enjoyment and the amenity of her life”. [142] His Honour accepted that she had a: [143]

His Honour took into account evidence of her pre-existing health, including episodic-specific anxiety and meningitis, which his Honour considered “did not have a significant effect on the plaintiff’s previous ability to fully exercise her earning capacity and enjoy the amenity of her life”. [144] His Honour also observed that her experience of having to undergo the evolving physical changes and restrictions associated with taking a pregnancy to full term, the ensuing caesarean section, and the subsequent post-operative infection, all comprised an impairment of her physical or mental condition and formed part of her damages. [145]

Observing that no comparable verdicts were cited, his Honour accepted the plaintiff’s submission that her non-economic loss should be assessed at 38% of a most extreme case. [146] This resulted in an award of the indexed amount of $261,000. On behalf of the doctor it was submitted that this was well outside the appropriate range.

The appellant emphasises that damages for non-economic loss are limited to those arising from physical injury or a recognised psychiatric disorder, and that grief, anxiety, distress and disappointment alone do not constitute a psychiatric disorder. [147] However, none of the trial judge’s findings, referred to above, as to the physical and mental impact on the patient of the pregnancy and infection was the subject of specific challenge on appeal. Nor was there any challenge to his Honour’s conclusion that the patient suffered from a recognisable psychiatric illness, namely a major depressive disorder, as was common ground between the relevant experts.

The appellant cited Tame v NSW, [148] for the proposition that it was not foreseeable that a person of normal fortitude would suffer a psychiatric injury from a false belief that she was incapable of becoming pregnant despite being warned that she could still become pregnant, and that any psychological reaction to her mistaken belief was neither foreseeable nor causally linked to any negligence on the part of the appellant. In my view, this is entirely unrealistic. The circumstances in Tame (in which it was held that it was not reasonably foreseeable that a person in the position of that plaintiff would sustain a recognisable psychiatric injury or illness as a result of the erroneous recording of her blood alcohol level in a motor accident report) is far removed from that of a failed tubal ligation, in which – accepting that the patient was told that there was nonetheless a risk of pregnancy of 1 in 2-300 – it was nonetheless quite foreseeable that she might become depressed upon discovering that she was in that unfortunate tiny minority.

Although no comparable cases were cited at first instance, the appellant referred to four in this Court:

As Basten JA explained in AEA Constructions Pty Ltd v Wharekawa, [153] an assessment of damages for non-economic loss under CLA, s 16 requires the court to determine the severity of a loss by reference to a proportion of a most extreme case, and in this respect:

Having regard to the comparable cases to which reference has been made, if I were myself assessing the patient’s non-economic loss, I might do so at about 33% of a most extreme case, regarding her psychological injuries as somewhat more severe than those in MD, but less than those in Sorbello. However, it does not follow that the trial judge’s assessment was not reasonably open. No particular error has been identified, and in the light of the 35% allowed in Sorbello, I do not consider that the 38% assessed in this case, though at the upper end, can be said to be outside the range of reasonableness.

Accordingly, this ground also fails.

My conclusions may be summarised as follows:

As to liability, the mere fact of failure, occurring as early as it did, was suggestive of operator error in application of the clip. That hypothesis was reinforced by the circumstance that the left clip was applied at a sub-optimal point on the tube and incorporated extraneous tissue, increasing the chances of an incomplete closure. The appearance of the clip in the HSG images did not exclude that hypothesis, because the clip could have been incompletely closed. The alternative explanation of recanalisation or fistula is rendered highly improbable by the relatively short time between operation and pregnancy, and by the evidence of Dr Jeri and the HSG, all of which militated against a fistula circumventing an otherwise successful transection. In those circumstances, the trial judge was right to conclude that more probably than not, the patient’s pregnancy post-tubal ligation was attributable to negligent operator error on the part of the doctor in the application of the left Filshie clip. Contrary to the appellant’s submissions, this approach involves no “reversal of the onus”.

As to damages, in proceedings involving a claim for the birth of a child, CLA s 71(1) does not preclude the award of damages for economic loss for loss of earnings by the claimant attributable to psychiatric injury associated with the birth of the child, as distinct from a need or choice to rear or maintain the child, and the trial judge’s award for economic loss was not erroneous. The assessment of non-economic loss at 38% of a most extreme case has not been shown to be affected by any particular error and, though perhaps at the upper end, cannot be said to be outside the range of reasonableness.

It follows that in my opinion the appeal should be dismissed with costs.

McCALLUM JA: I agree with the orders proposed by Brereton JA for the reasons his Honour has comprehensively and persuasively explained.

SIMPSON AJA: I agree with Brereton JA.