Gouldsmith v Mid Staffordshire General Hospitals NHS Trust

This is an appeal against the judgment of His Honour Judge Gilmour QC given on 19 December 2005. The judge dismissed a claim by Mrs Irene Gouldsmith (“the appellant”) for damages for personal injuries sustained as a result of alleged negligence while she was under the care of the Mid Staffordshire General Hospital NHS Trust at Stafford General Hospital in 1999. The trial was on liability, the question of damages being deferred. The judge found that the respondents had been negligent but that their negligence had not caused the injury suffered. The injury was the loss of the digits of her left, dominant, hand. The left index finger was amputated on 1 September 1999 and the remaining digits on 22 October.

The appellant’s case was that she should have been referred to a specialist, tertiary, hospital when problems with her left hand were encountered. Had she been so referred, appropriate surgery would have been conducted which would have saved the digits. The judge held (paragraph 38 of his judgment) that, on balance of probabilities, an operation by 13 October 1999 would have resulted in the saving of the second and third fingers and the thumb, though not the index finger.

The appellant was referred to the Stafford General Hospital on 14 August. It is common ground that she suffered from arterial thrombosis (a tendency to clotting of the blood) and antiphospholipid syndrome (“AS”) which gave rise to a tendency for vessels to become narrowed or blocked by blood clots. She presented with an ischemic left finger. It is also common ground that the appellant also had a lesion, a stenosis just beyond the origin of the left subclavian artery.

The judge set out the history of treatment, the issues, and the submission of the parties, with great care and clarity. At paragraph 5, he summarised the appellant’s case.

“At the time of presentation at the Defendant’s hospital on 14 August it is agreed that the Claimant had what I will hereafter call “a lesion”, namely a stenosis just beyond the origin of the left subclavian artery together with blockage of both the ulnar and radial arteries. It is the Claimant’s case that that lesion was responsible in whole or in part for a deterioration in the condition of the left dominant hand which led initially to partial amputation of the left index finger and ultimately on the 22 October with amputation of the remaining digits. It is the Claimant’s case that the stenosis should have been dealt with by operative procedure and if the Defendants did not have the medical staff qualified to carry out what would be a complicated and difficult procedure, the Claimant should have been referred to a centre of excellence where it is said she would have received an operation to remove the lesion and she would have avoided the amputations that occurred on 22 October. The essence of the Claimant’s case was set out in paragraph 5 of the opening argument. “It is the essence of the Claimant’s case … that had appropriate intervention in relation to the source of the embolisation taken place such amputations would not have occurred.”

The respondents’ case was set out at paragraph 6:

“It is the Defendant’s case that although the lesion may have produced emboli ( a piece of solid material in the blood circulation) causing the symptoms that brought the Claimant to hospital in August, her deterioration is not shown to be caused by further emboli from that source. The Defendants contend that it is equally or more likely to be the underlying tendency to thrombosis that caused the deterioration. Therefore they argue that an operation on the lesion would serve no purpose because that was not the cause of her deterioration. The only treatment that might help was the treatment she received namely anticoagulation and the reason that did not work was possibly due to her non compliance.”

The judge noted that two distinguished professors of surgery were unable to agree as to what was the cause of the deterioration in the condition of the left hand that led to the amputation of the digits on 22 October. He stated the issue at paragraph 7:

“The doctors are agreed that the small vessels in the Claimant’s left hand and arm were occluded by thrombus or embolic material that was not fresh. Professor McCollum believes that these occlusions were due to embolism from an exuberant atherosclerotic plaque or adherent thrombus in the left subclavian artery and that these emboli have caused secondary thrombosis involving the small vessels in which emboli lodged. Professor Bell is of the opinion that AS caused primary thrombosis in the small vessels of the forearm and hand.”

The appellant was in the care of Mr Durrans, consultant surgeon with an interest in vascular matters, who first saw her on 16 August. An angiogram was carried out by Dr Willard, consultant radiologist, on 17 August. Dr Willard found a large exuberant plaque causing stenosis just beyond the origin of the left subclavian artery. His conclusion was:

“Overall the appearances are consistent with distal embolisation from the subclavian stenosis - - the subclavian stenosis could be stented at a later date.”

Dr Willard’s view was that the blockages of the vessels of the forearm could be explained with emboli breaking off from the lesion.

Management of the appellant included the use of anti-coagulants. There were periods of in-patient treatment. Symptoms began in the left index finger but later developed in other digits. The appellant did not fully co-operate with treatment but there is no allegation of contributory negligence. There were periods of improvement but these were followed by further deterioration in the condition of the hand.

The appellant was re-admitted to hospital on 1 October but was discharged on 7 October to be seen in the clinic in two weeks time. The appellant next attended the hospital on 13 October when there was obvious evidence of deterioration in the left hand. By 18 October the appellant was complaining of great pain and the condition of the digits was worsening. The possibility of amputation of the remaining four digits was discussed and was conducted on 21 October.

The judge assessed the evidence of the two main expert witnesses. He found that they were both very well qualified by experience and medical expertise. Professor McCollum, Professor of Surgery, University of Manchester, was a fair witness seeking to be helpful: “Overall I thought that he was an impressive witness although this does not mean that all his opinions were necessarily correct”. Professor Bell, Emeritus Professor of Surgery, University of Leicester, was in a number of respects, he said, an unsatisfactory witness. That assessment is justified by examples given. The judge stated that he did not doubt that Professor Bell was giving his true medical opinion when he said that the clavicular lesion was “not to be seen as playing a part in the deterioration in the left hand”. Professor Bell’s evidence was that he would not have carried out any operation to the lesion in August or September 1999 but would have adopted essentially the same treatment as that adopted by Mr Durrans. The clavicular lesion was not the source of emboli causing the amputation, in his opinion.

In the course of closing submissions at the trial by Mr Feeny, for the respondents then as now, the judge stated that he found Professor McCollum, “one of the most impressive medical witnesses I have ever heard. I thought he was logical. He gave the appearance of being fair and I have to say that my note at the time when he finished his evidence was that he was a very impressive witness”. While not prepared to accede to the invitation of Mr Limb, for the appellant, that he would find for Professor McCollum on each point of disagreement, the judge confirmed that he did not go back on what he had said about Professor McCollum. Other medical witnesses, Dr Hay, Dr Bruce and Professor Jayson were found by the judge to be careful witnesses who were seeking to be helpful.

At paragraph 21, the judge again stated the central issue in the case:

“Was the deterioration in the appellant’s dominant hand between August and October 1999 caused or contributed to by continuing embolisation from the subclavical lesion?”

While it was agreed that an embolism had been shown on the angiogram on 17 August, there was an issue as to whether continuing embolisation had occurred. The judge set out Professor McCollum’s reasons for believing that it had: no previous vascularisation of the digits notwithstanding the disorder present; no other features of spontaneous thrombosis; it was the limb with the lesion that presented the continuing problem; the “patchy filling” of the distal vessels shown on angiogram was more likely to result from embolism than from the disorder.

Dr Hay also favoured embolisation from the lesion as the cause of deterioration, relying on the fact that there had been three blocked arteries in a short period of time. He thought embolisation more likely to be responsible for the deterioration than primary thrombosis. Professor Jayson stated that he had great difficulty understanding the mechanism of the respondents’ alternative diagnosis, micro embolisms blocking the ulnar and radial arteries.

In the opinion Professor Bell, the AS syndrome was on the balance of probabilities the most likely cause. Reliance was placed on the underlying tendency to thrombus, which condition had subsequently advanced to involve the patient’s legs. Professor Bell said that in his long medical career he had not seen the sequence of events suggested by the appellant’s experts. The underlying blood disorder can explain the deterioration in the hand and there was no need to look further for a cause.

At paragraph 21, the judge stated:

“I accept the Claimant’s argument that the contemporaneous notes are consistent with or positively indicative of the view that the ischaemia in the fingers was connected with embolisation from the sub-clavian stenosis.”

The judge then said how difficult he found his decision. At pargraph 27, he stated:

“On a fine balance I find that there was some continuing embolisation from the lesion into the vessels of the left hand that played a causative role, of uncertain amount, in the deterioration of the left hand … All I am able to say on the balance of probabilities is that continuing embolisation from the lesion played a causative part in the loss of the digits.”

On those findings, the judge considered the question of breach of duty. He referred, at paragraph 30, to the appellant’s case:

“Mr Durrans was a general surgeon albeit with a special interest in vascular surgery. Initially it was perfectly appropriate for him to rely on anticoagulation but there came a time when it should have been obvious to Mr Durrans that anticoagulation was not working. There came a time when he was faced with a patient whose dominant limb is deteriorating and the possibility of loss of digits arises. According to Professor McCollum if there is only a 10% chance of saving the dominant hand then you are obliged to seriously consider an operation if that will improve matters.”

The judge referred to the evidence of Professor McCollum:

“We are heading toward a major amputation of the limb and the things that they thought about were not done. If you were getting out of your depth the Consultant Surgeon has a duty to refer a patient to a tertiary centre where they may have more experience with this type of case.”

The judge held that in August and September, the appellant could reasonably be seen as having a rare and unusual vascular condition. Having referred to Mr Durrans’s evidence, the judge concluded, at paragraph 33:

“In my view Mr Durrans was negligent in not referring the Claimant to an appropriate centre of excellence wherever that may have been. There came a time (which I shall discuss in a moment) when he ought to have appreciated that he had exhausted his knowledge and expertise. Anticoagulation was not working and it seems to me it does not particularly matter why it is not working. He was faced with a patient who was unwilling or unable to comply with the taking of anticoagulation. Mr Durrans would not be entitled to simply say “it’s her fault there is nothing more that I can do”. Quite simply Mr Durrans should have concluded, by the date that I will decide in a moment, that the best chance of saving the patient’s hand was to refer to a specialist centre.”

The judge went on to find that during the first week of October, Mr Durrans should have referred the patient elsewhere and his omission to do so was negligent.

In a respondents’ notice, Mr Feeny has challenged the judge’s finding that there was a continuing embolism following the diagnosis of 17 August. The judge plainly relied on the evidence of Professor McCollum and we have been referred to transcripts showing the thorough testing of his evidence on that issue. The view was cogently and consistently expressed by Professor McCollum that the embolism had continued and I am in no doubt that the judge was entitled to rely on his evidence.

The second point in the respondents’ notice refers to dates. It is submitted that the judge was not entitled, on the evidence, to hold that 13 October 1999 was a critical date in terms of saving the digits. First, it was not the date advocated by Professor McCollum in his oral evidence, in which he relied on an earlier date, and second, there is not sufficient evidence of a gradual deterioration in the condition to establish a date by which the digits could have been saved. I do not accept those submissions. While this was not a condition in which the rate of deterioration was consistent, Professor McCollum stated, in oral evidence:

“What happens to patients who have a proximal source of embolisation that is not removed, a continued deterioration is expected in those patients.”

Further, the experts agreed in their joint report of September 2004:

“The remaining fingers were severely painful and the vascular experts agree that these fingers may still have been viable when they were amputated [that is on 22 October]”.

On the evidence, the judge was entitled to find that surgery by 13 October 1999 would have resulted in the saving of the second and third fingers, and the thumb, of the left hand.

The judge’s finding on breach of duty is not further challenged in this appeal. The judge had every opportunity, with the help of skilful cross-examination, to consider the medical evidence and his conclusion on that issue is not in my judgment vulnerable to challenge.

However, the judge went on to find that the appellant’s injury had not been caused by the breach of duty. He accepted that the appellant would have been treated as an urgent case upon reference but continued, at paragraph 36(c):

“I am prepared to infer that the Claimant would have undergone a range of tests that would have revealed the connective tissue disorder from which the Claimant suffered. But would the new doctors have formed the view that the lesion had to be operated upon? In the absence of evidence on this point I cannot say on the balance of probability that the Claimant would have had the lesion operated upon because I do not know the identity of the hospital or the surgeon who would have seen her. Can I be satisfied on the balance of probability that every specialist exercising the proper standard of care would have operated on the Claimant? The answer to this is no. It is just as likely that a vascular surgeon would in accordance with a proper standard of care have formed the same view as Professor Bell namely that nothing would be served by operating upon the lesion. I cannot assume that the new vascular surgeon with support of a haematologist would have advised an operation. The highest the case can be put on this point from the Claimant’s point of view is the evidence of Professor McCollum at page 7.80 of the bundles namely he “considered that most specialist vascular units would have undertaken surgery to prevent further embolisation from the obvious lesion in the left subclavian artery”. This evidence is not sufficient to allow me to conclude on the balance of probability that any unit to which the Claimant might have been referred would have been bound (in the Bolam sense) to have operated.

Accordingly this claim fails on the basis that I am unable to conclude on all the available evidence that had the Claimant been referred to a specialist vascular unit in early October she would have received the treatment which, according to her, would have saved her digits. ”

Mr Limb’s main submission is that the judge has in these paragraphs applied the wrong test. The judge has not first considered, as he should have done, what, on balance of probabilities, would in fact have happened upon a reference to a specialist hospital but the different question whether the authorities there “would have been bound (in the Bolam sense) to have operated”. However, I first consider his submission that, even applying the Bolam test, set out in paragraph 26 below, the judge should have found for the appellant on causation.

In support of that submission, Mr Limb relies on the judge’s acceptance of Professor McCollum’s evidence, which included the statement that it would have been “bizarre” for the respondents to ignore the subclavical lesion, which was ‘very dramatic’. It followed from that and from the evidence that the respondents’ doctors were “getting out of [their] depth” that an operation would plainly have followed referral. Once Professor Bell’s evidence that there was no continuing embolism was rejected, it was inevitable that surgery would have been conducted at the specialist centre to which the patient should have been referred. Professor Bell had accepted that if, contrary to his view, the patient had an embolism, an operation would have helped. Professor McCollum had expressed the view that the failure to treat the stenosis at the origin of the left subclavian artery could not be defended. By the beginning of October, it was “spectacularly obvious” that something should have been done. A second opinion was required while the limb was still viable.

A continuing deterioration in the condition of a dominant limb known to have been subject to recent embolism would appear to be a very strong candidate for more radical treatment than continued anti-coagulation. The judge’s rejection of Professor Bell’s diagnosis gives weight to Mr Limb’s submission, but, I am not, having regard to the judge’s careful consideration of the evidence, able to reverse the findings of the judge on that point. He did not reject Professor Bell’s opinions as being without foundation. It would not be appropriate for this court to substitute a conclusion that the surgery Professor McCollum considered urgent and essential must inevitably have been conducted at the hospital of reference.

I turn to the main submission. In my judgment, the judge has, in his analysis, not applied the correct test. The approach to be followed in circumstances such as the present was considered by the House of Lords in Bolitho v City and Hackney Health Authority [1998] AC 232. At p239C, Lord Browne-Wilkinson first cited the Bolam test:

“The locus classicus of the test for the standard of care required of a doctor or any other person professing some skill or competence is the direction to the jury given by McNair J. in Bolam v Friern Hospital Management Committee [1957] 1 W.L.R. 583, 587:

‘I myself would prefer to put it this way, that he is not guilty of negligence if he has acted in accordance with a practice accepted as proper by a responsible body of medical men skilled in that particular art … Putting it the other way round, a man is not negligent, if he is acting in accordance with such a practice, merely because there is a body of opinion who would take a contrary view.’

It was this test which Lord Scarman was repeating, in different words, in Maynard’s case in the passage by reference to which the judge directed himself.”

Lord Browne-Wilkinson, with whom Lord Slynn of Hadley, Lord Nolan, Lord Hoffmann and Lord Clyde agreed, then stated, at p239F to 240F, the correct test or tests to be applied when causation is in issue:

“Where, as in the present case, a breach of a duty of care is proved or admitted, the burden still lies on the plaintiff to prove that such breach caused the injury suffered: Bonnington Castings Ltd. v. Wardlaw [1956] A.C.613; Wilsher v. Essex Area Health Authority [1988] A.C. 1074. In all cases the primary question is one of fact: did the wrongful act cause the injury? But in cases where the breach of duty consists of an omission to do an act which ought to be done (e.g. the failure by a doctor to attend) that factual inquiry is, by definition, in the realms of hypothesis. The question is what would have happened if an event which by definition did not occur had occurred. In a case of non-attendance by a doctor, there may be cases in which there is a doubt as to which doctor would have attended if the duty had been fulfilled. But in this case there was no doubt: if the duty had been carried out it would have either been Dr. Horn or Dr. Rodger, the only two doctors at St. Bartholomew's who had responsibility for Patrick and were on duty. Therefore in the present case, the first relevant question is 'What would Dr. Horn or Dr. Rodger have done if they had attended?' As to Dr. Horn, the judge accepted her evidence that she would not have intubated. By inference, although not expressly, the judge must have accepted that Dr. Rodger also would not have intubated: as a senior house officer she would not have intubated without the approval of her senior registrar, Dr.Horn.

Therefore the Bolam test had no part to play in determining the first question, viz. what would have happened? Nor can I see any circumstances in which the Bolam test could be relevant to such a question.

However in the present case the answer to the question 'What would have happened?' is not determinative of the issue of causation. At the trial the defendants accepted that if the professional standard of care required any doctor who attended to intubate Patrick, Patrick's claim must succeed. Dr. Horn could not escape liability by proving that she would have failed to take the course which any competent doctor would have adopted. A defendant cannot escape liability by saying that the damage would have occurred in any event because he would have committed some other breach of duty thereafter. I have no doubt that this concession was rightly made by the defendants. But there is some difficulty in analysing why it was correct. I adopt the analysis of Hobhouse L.J. inJoyce v. Merton, Sutton and Wandsworth Health Authority [1996] 7 Med.L.R. 1. In commenting on the decision of the Court of Appeal in the present case, he said, at p. 20:

'Thus a plaintiff can discharge the burden of proof on causation by satisfying the court either that the relevant person would in fact have taken the requisite action (although she would not have been at fault if she had not) or that the proper discharge of the relevant person's duty towards the plaintiff required that she take that action. The former alternative calls for no explanation since it is simply the factual proof of the causative effect of the original fault. The latter is slightly more sophisticated: it involves the factual situation that the original fault did not itself cause the injury but that this was because there would have been some further fault on the part of the defendants; the plaintiff proves his case by proving that his injuries would have been avoided if proper care had continued to be taken. In the Bolitho case the plaintiff had to prove that the continuing exercise of proper care would have resulted in his being intubated.'

There were, therefore, two questions for the judge to decide on causation. (1) What would Dr. Horn have done, or authorised to be done, if she had attended Patrick? And (2) if she would not have intubated, would that have been negligent? The Bolam test has no relevance to the first of those questions but is central to the second.”

Those were the appropriate questions in Bolitho because the negligence was in the failure of the hospital doctor, Dr Horn, to attend a patient in the hospital. It was necessary to consider whether that failure was causative of the damage. Dr Horn’s evidence was that, had she attended, she would not have intubated. That being so, the second question posed by Lord Browne-Wilkinson at p240G arose: “Would it have been negligent not to have intubated?”

In the present case, the breach of duty was the failure to refer to a specialist hospital and the first question on causation is: “What would have happened upon reference to a specialist hospital?”, the first question posed by Lord Browne-Wilkinson in Bolitho. If the answer to the question is that appropriate surgery would probably have been conducted, the second question: “Would it have been negligent not to operate?” does not arise. Unlike the situation arising on the facts in Bolitho, that answer to the question “What would have happened?” is determinative of the issue of causation. As Lord Browne-Wilkinson stated, at p240B, the Bolam test has no part to play in determining the first question, “viz. what would have happened?” The situation in Bolitho was different in that Dr Horn gave evidence that, had she attended, she would not have intubated and, that being so, the second question posed in Bolitho arose. It is necessary to apply that analysis to the judge’s reasoning in the present case.

The judge did not follow the correct sequence. Read alone, it could be argued that the second and third sentences of the judge’s findings in paragraph 36(c), cited at paragraph 22 above, were directed to Lord Browne-Wilkinson’s first question. On a reading of the paragraph as a whole, it appears to me that the judge allowed the first question to be subsumed within the second question: “Would every specialist exercising the proper standard of care have operated?” Professor McCollum’s evidence is addressed, later in the paragraph, not with a view to answering the first question: “What would have happened?, as it should have been, but the question “Would the specialist unit have been bound (in the Bolam sense) to have operated?” The paragraph cannot be read as addressing the Bolitho questions consecutively.

If the opinion of Professor McCollum cited by the judge was accepted, as it appears to have been, the second question did not arise. Had the patient been referred, as she should have been, the saving operation would on balance of probabilities have been conducted. That is, as Hobhouse LJ put it in Joyce; “The factual proof of the causative effect of the original fault”. The appellant does not need to travel the alternative route considered by Hobhouse LJ, that is, proof that the hospital of reference would have been bound (in the Bolam sense) to have operated.

I add that the passage cited by the judge from the evidence of Professor McCollum, an expert witness for whose evidence he had the highest regard, was not the only evidence from Professor McCollum on this point. He had stated in his letter of 25 July 2005;

“It is also my opinion, on the balance of probabilities, that had she been referred to a Specialist Unit (usually in a teaching hospital), then the balance of probabilities are that a procedure would have been done within a few days of arrival to prevent further embolisation from the subclavian artery lesion. In my opinion, although it would have been too late to save the left index finger, such a procedure would have prevented subsequent trans-metatarsal amputation of the left hand”.

In my judgment, the conclusion of the judge on causation cannot stand when the test has not been applied correctly. It has not been urged upon the court that, in that event, there a should be a re-trial. I do not find that surprising when the judge has referred on causation to the evidence of Professor McCollum, and only his evidence, which, if applied to the right question, concludes the case on causation in the appellant’s favour. Mr Feeny did, however, argue that the judge should not have accepted the appellant’s evidence on what would have happened in fact because there was contrary evidence from Professor Bell.

Professor Bell did not address this issue directly. His evidence throughout was that the anti-coagulant treatment adopted by Mr Durrans was the appropriate treatment, and that no reference was appropriate. The issue of what could have happened upon a reference was not directly addressed. The respondents’ evidence was directed to justify the adequacy of anti-coagulant treatment. It is not appropriate to allow evidence given, and not accepted, to justify non-referral, into a world in which referral has been found to be necessary.

Reference has been made to Professor Bell’s statement, in the joint report of July 2005. He stated:

“PB does not think surgery was indicated as the forearm vessels were already blocked and surgery would have been dangerous. He is therefore of the opinion that most specialist vascular units would not have operated under these circumstances”.

That is the nearest Professor Bell came to dealing with a world in which referral had hypothetically occurred. What he did, however, was to re-assert his opinion that the problem was in the “forearm vessels”. It was because of that (“therefore”) that most specialist vascular units would not have operated. Upon the judge’s findings that the problem was elsewhere, that evidence cannot be used, and the judge did not attempt to use it, to defeat the opinion of Professor McCollum as to what would probably have happened on referral. Had the need for referral been recognised, it would have been upon the basis that the lesion in the subclavian artery was potentially the source of the problem and that would have been recognised in the letter of reference. That is not a situation which Professor Bell has been prepared to contemplate, believing that referral was unnecessary, the lesion in the subclavicular region being irrelevant. He also stated that “if the patient had an embolus, which she did not (my emphasis), then thrombolysis or even an operation would have helped.” (The second of those propositions actually supports the appellant’s case on causation once the correct diagnosis is made).

Though not proposed by the respondents, I have considered the possibility of remitting the issue of causation for re-trial. It would, in my judgment, be unjust to do so. It would give the respondents a second opportunity to call evidence on causation when, at the trial, their case was entirely directed to contesting the breach of duty and did not directly contest a finding of causation if their primary case failed. Further, the judge cited, and appears to have accepted, evidence of Professor McCollum which concludes the issue of causation in the appellant’s favour.

I consider that the judge should and would have found, upon a correct application of the test, that the breach of duty was causative of the damage. It would be unfair to permit the respondents to have a further opportunity to call evidence on the point when their entire case has been based on the adequacy of anti-coagulant treatment at the respondents’ hospital. That argument having been rejected by the judge, along with Professor Bell’s refusal to accept the relevance of the subclavicular lesion, Professor McCollum’s evidence as to what would have happened on referral should be accepted. This is not weakened by evidence as to the complexity of the operative procedure concerned. Professor Bell’s entire case has been based on a premise which cannot, in the light of the findings of the judge, be sustained when considering what future treatment was likely.

Counsel were invited to comment on the majority decision of the House of Lords in Gregg v Scott [2005] 2 AC 176 where a different question arose on the issue of causation in medical negligence cases. They did not consider it relevant to the present issue and I agree. On a finding, on balance of probabilities, that the negligent failure to refer was the cause of the injury complained of, consideration of the loss of a chance does not arise. No adverse comment was made in Gregg about Bolitho and it is the tests propounded in Bolitho case which apply to the present facts.

For those reasons, I would uphold the judge’s conclusion on breach of duty, allow the appeal on causation and remit the issue of damages to the High Court.

I regret to say that I have come to a different conclusion to that reached by Pill and Wilson LJJ. The crucial passages in the judgment of the trial judge are in paragraph 36(c) and paragraph 37. It is appropriate to set them out again:

“… Would the new doctors have formed the view that the lesion had to be operated upon? In the absence of evidence upon this point I cannot say on the balance of probability that the claimant would have had the lesion operated upon because I do not know the identity of the hospital or the surgeon who would have seen her. Can I be satisfied on the balance of probability that every specialist exercising the proper standard of care would have operated upon the claimant? The answer to this is no. It is just as likely that a vascular surgeon would in accordance with a proper standard of care have formed the same view as Professor Bell, namely that nothing would be served by operating on the lesion. I cannot assume that the new vascular surgeon with support of a haematologist would have advised an operation. The highest the case can be put on this point from the claimant’s point of view is the evidence of Professor McCollum … namely he ‘considered that most specialist vascular units would have undertaken surgery to prevent further embolisation from the obvious lesion in the left subclavian artery’. This evidence is not sufficient to allow me to conclude on the balance of probability that any unit to which the claimant might have been referred would have been bound (in the Bolam sense) to have operated.”

I accept that paragraph 36(c) falls below the level of clarity that characterises the rest of the judgment. However, in my judgment, properly deconstructed its reasoning is as follows:

That is the sense in which I read paragraph 36(c). I do not consider it to be a strained reading. Nor do I think that my interpretation is the result of my rewriting the judgment. Moreover, such findings cannot be said to do any injustice to the totality of the evidence.

The next question is as to the legal test to be applied to those findings. The starting point is Bolitho v City and Hackney Health Authority [1998] AC 232, in which Lord Browne-Wilkinson adopted (at p.240D) the following analysis by Hobhouse LJ in Joyce v Merton, Sutton and Wandsworth Health Authority (1996) 7 Med LR 1, at p.20:

“Thus a plaintiff can discharge the burden of proof on causation by satisfying the court either that the relevant person would in fact have taken the requisite action (although she would not have been at fault if she had not done so) or that the proper discharge of the relevant person’s duty towards the plaintiff required that she take that action.”

These are alternative routes to proving causation. Lord Browne-Wilkinson then observed that the Bolam test has no relevance to the former but is central to the latter (p.240G).

In the present case, the first route requires proof on a balance of probabilities of what would have happened if there had been a referral to a centre of excellence. Pill LJ takes the view that the best evidence on this issue was that referred to by the judge, namely Professor McCollum’s opinion that “most specialist vascular units would have undertaken surgery to prevent further embolisation …” That, it is said, enabled the appellant to succeed on the balance of probabilities.

Even without resort to the evidence of Professor Bell, I do not consider that those words of Professor McCollum get the appellant home. The statement that “most” would take a certain course carries with it the clear implication that some would not. In Bolitho the factual parameters were narrow because the actual and the hypothetical facts all occurred within the same hospital, the identity of the hypothetical actor was apparent and the evidence gave rise to a finding of fact that, if she had been there and had considered whether or not to intubate, her decision would have been not to do so. Where, as in the present case, the negligence takes the form of a failure to refer to a different, more expert unit in a different city, there is an evidential vacuum unless the claimant fills it by establishing (1) where she would probably have been sent and (2) what the probable treatment would have been at that destination. This appellant did not fill that vacuum. There were at least two possible destinations and there was no specific evidence about the hypothetical approach at either.

I am not persuaded that, in the light of Professor McCollum’s evidence, the judge should have resolved this issue in favour of the appellant. The implication is that the respondent was remiss in not calling contrary evidence. In my judgment, that not only disregards the burden of proof; it also ascribes to Professor McCollum’s evidence on this point more probative force than its terms naturally convey. This appears to have been the view of the judge. I consider it to have been a permissible one. The difficulties in this case (and it was always a difficult case) were intensified when the appellant’s primary case (that there should have been surgical intervention by the respondent’s team) and the respondent’s primary case (that there was no continuing embolisation) both unravelled in the course of the trial with the result that the eventual crunch issues of failure to refer and its consequences were never canvassed in evidence as extensively as they would have been if it had been appreciated that they would assume the significance that they eventually did. Such things happen in this sort of litigation. When they do, the greater disadvantage will often fall upon the party with whom the burden of proof resides. In my view, that is what happened here. Not anticipating that the case would turn on the failure to refer and its consequences, the appellant did not address specific evidence about the probable course of the particular hypothetical referral.

In Bolitho, Lord Browne-Wilkinson recognised the distinctive feature of that case when he said (at p.239H) that “there may be cases in which there is a doubt as to which doctor would have attended if the duty had been filled. But in this case there was no doubt”. That was a case of non-attendance within a single hospital. Where the issue is non-referral to an unspecified different hospital, the scope for doubt is that much greater.

The essential difference between my analysis and that of Pill and Wilson LJJ is that I do not consider that the judge applied the wrong test. In a case in which there is no established probability about where or by which consultant the decision to operate would have been considered, the first question (what would have happened?) has not been answered on a balance of probability because, on Professor McCollum’s evidence, in some places and in the hands of some consultants, a decision not to operate could reasonably eventuate. Even where, on the accepted evidence, most consultants in most centres of excellence would have decided upon surgical intervention, a claimant has not discharged the burden of proof if he or she has failed to establish that the hospital to which and the consultant to whom the reference would have been made would probably have been one of the majority rather than one of the still Bolam-compliant minority. In this regard, factual circumstances involving hypothetical reference elsewhere put different obstacles in the way of a claimant when compared with the entirely “in-house” circumstances of Bolitho.

I do not underestimate the difficulty faced by a claimant who has to prove, on a balance of probability, which of two non-negligent courses would have been adopted by a non-party. It seems to me that it was such difficulty that lead Lord Hoffmann in Gregg v Scott (at para 83) to opine that where it is the hypothetical response of a non-party, rather than the act of the claimant, the defendant or someone for whom the defendant is responsible, that is in issue, the question should be approached on the basis of loss of a chance rather than proof on a balance of probability. However, that is a road down which neither counsel in the present appeal, and neither of the other members of the court, wish to travel on this occasion. Accordingly, I shall say no more about it save to observe that it has some eminent academic support: see Stapleton, Cause-in-fact and the Scope of Liability for Consequences (2003) 119 LQR 388, 403-404.

For the reasons I have given, I find no fault in the reasoning of the trial judge and, for my part, I would dismiss this appeal.

I agree with the judgment of Pill L.J.

In his final written submissions to the judge Mr Feeny addressed as follows the issue upon which the appellant’s claim was to founder:

“… the Claimant has to prove causation in terms of surgery at a different centre. There is no factual evidence on this issue and therefore the Claimant must show that all competent specialist surgeons would have operated in these circumstances.”

Mr Feeny’s submission duly reflected the analysis of the House of Lords in Bolitho v. City and Hackney Health Authority [1998] AC 232. That was a case in which a doctor was in breach of duty in failing to give medical attention to a patient. The present is a case in which a doctor was in breach of duty in failing to secure other, specialist, medical attention for a patient. There is no relevant difference. So the two questions articulated by Lord Browne-Wilkinson in Bolitho at 240 B-G are directly in point. Framed around the facts of the present case, they are:

My view is that the judge proceeded down the path which Mr Feeny had set before him. In his judgment he said, at §36(c),:

“But would the new doctors have formed the view that the lesion had to be operated upon? In the absence of evidence on this point I cannot say on the balance of probability that the Claimant would have had the lesion operated upon because I do not know the identity of the hospital or the surgeon who would have seen her. Can I be satisfied on the balance of probability that every specialist exercising the proper standard of care would have operated upon the Claimant? …”

Thereupon the judge proceeded to give a negative answer to the second question which he had there posed; and in my view the whole of the rest of the sub-paragraph, set out by my colleagues at §22 and §40 above, contains the judge’s explanation for that negative answer.

So, while I consider that Mr Feeny was wise to concede before us that paragraph 36(c) of the judgment is “difficult”, I accept that, albeit very briefly, the judge did address the first question. My view, however, like that of Pill L.J., is that, by reference to evidence which he accepted, the judge should have given it an affirmative answer. For Mr Feeny had been wrong to submit to him that there was no factual evidence on the first question. As I will explain, there was such evidence; and the irony is that the judge not only considered but accepted it when he addressed the second question.

There was no obligation on the respondents at any stage to volunteer evidence as to the identity of the specialist unit to which Mr Durrans would have been likely to refer the appellant if, which they denied, it had been his duty to make such a referral. The fact is, however, that they never did volunteer it. Mr Durrans did not address the point in his evidence in chief; nor did Mr Limb, whose strategy was different, ask him about it in cross-examination. Thus in §36(a) of his judgment the judge correctly observed that there was no evidence of the identity of the hospital to which the appellant would have been referred. In particular there had in my view been no “evidence” in the proper sense that the reference would have been to Birmingham or to Manchester. The only reference at the hearing to Birmingham and Manchester occurred when, in response only to a specific invitation from the judge for assistance in this regard, Mr Limb asked Professor Bell during cross-examination whether he could comment on his instructions that the reference would be likely to have been to units either in Birmingham or in Manchester. In the event the professor expressed himself doubtful whether either unit had the necessary expertise. Nevertheless it was by reference to that nebulous exchange that in §36a the judge added that, notwithstanding the absence of evidence, he was prepared to “assume” that the reference would have been either to Birmingham or to Manchester.

During the assembly of the case for trial there was therefore no suggested identification of a specialist unit in relation to which the first question articulated in Bolitho could be explored. But did it follow that the question could not be answered at all? The parties clearly – and in my view correctly – considered otherwise. They agreed that the two vascular experts should be asked to give a joint response to the following written question, precisely reflective of the first question articulated in Bolitho:

“Had the Claimant been referred to a specialist vascular unit, on the balance of probabilities, would she have been operated on?”

Indeed on 16 February 2005 a district judge ordered the experts to answer the question.

In their joint written response the experts gave different answers to the question. But they both chose to answer it by reference to the same criterion, namely by reference to what “most” specialist units would have done. I suspect that it was almost instinctive for them to answer it in that way; at all events such was in my view the logical approach for them to take. They wrote:

“[Professor McCollum] considered that most specialist vascular units would have undertaken surgery to prevent further embolisation from the obvious lesion in [the appellant’s] left subclavian artery. [Professor Bell] does not think surgery was indicated as the forearm vessels were already blocked and surgery would have been dangerous. He is therefore of the opinion that most specialist vascular units would not have operated under these circumstances.”

As Pill L.J. has explained at §35 above, the judge clearly rejected Professor Bell’s opinion that most units would not have operated; for the opinion was predicated upon an assumption that surgery would have been pointless, which, in finding that it would be likely to have saved two fingers and a thumb, the judge rejected. What was left, therefore, was Professor McCollum’s contrary opinion. Study of the judge’s analysis of the second question articulated in Bolitho in §36(c) of his judgment in my view shows that he expressly accepted Professor McCollum’s contrary opinion. For, after describing it as “[t]he highest the case can be put on this point from the Claimant’s point of view” and after then quoting it, the judge said only:

“This evidence is not sufficient to allow me to conclude on the balance of probability that any unit to which the Claimant might have been referred would have been bound (in the Bolam sense) to have operated.”

Thus the judge did not say that he did not accept Professor McCollum’s opinion. On the contrary he took it as the high point of the claimant’s case and accepted it. His conclusion, in my view entirely correct, was only that the opinion did not enable him to give an affirmative answer to the second question articulated in Bolitho. But what, lulled by Mr Feeny’s silver tongue, the judge had overlooked was that the opinion had been specifically obtained by the parties in relation to the first question and was directly relevant thereto.

In the end Mr Feeny submits to us that, in order to answer the first question as to whether the specialist was likely on the balance of probabilities to have operated upon her, it was insufficient for the appellant to prove that most specialists would be likely to have done so. He first submits that, if most would be likely to have done so, it follows that some would not be likely to have done so. In my view that submission pays insufficient respect to the court’s task to balance probabilities. He then submits that it is for the claimant somehow to collect the evidence which would identify the unit to which she would have been referred, and indeed presumably the specialist consultant within that unit into whose charge she would have been placed, and to prove that he or she would have operated upon her. With great respect to Maurice Kay L.J., who accepts it, I regard this submission as placing a heavy burden upon a claimant for which there is no logical justification. We should not forget the reasons why in this case the situation has arisen in which the identity of the specialist is unclear. The first is that, in breach of the duty of care, Mr Durrans failed to refer the appellant to a specialist and thereby banished the parties to these realms of speculation. The second is that, in relation to the likely choice of specialist, being a matter which is peculiarly within their knowledge rather than that of the appellant or even of her experts, the respondents have elected to volunteer no information. The effect of Mr Feeny’s submission is that the respondents could wait to see whether the appellant managed to collect evidence in support of an asserted identification of the likely specialist; could wait to see whether such specialist, if still alive and available, asserted that he or she would have been likely to have operated upon the appellant; and could then seek to rebut either or both of such assertions. I reject the submission. The fact established by the appellant, namely that most specialists would be likely to have operated upon her, prima facie justified the conclusion that the specialist to whom the respondents should have referred her would be likely to have done so. It was a conclusion which, in the absence of any other evidence which cast doubt upon it, should have prevailed: in other words her establishment of the fact shifted the evidential burden of proof on to the respondents. It was open to them to have countered, had they had the material with which to do so, with evidence that the reference would be likely to have been to a particular specialist who would not have operated on the appellant. In the absence of credible evidence of that character the answer to the first question proffered on behalf of the appellant should have secured judgment for her.