Roughton v Weston Area Health Authority & Ors

This is an appeal against the decision of Mr Justice Owen in a judgment dated 17 November 2003 to dismiss a claim for damages by Adelaide Roughton (‘the Claimant’) against general medical practitioners and a health trust. In April 1998, when she was 35 years old, the Claimant suffered an acute and severe episode of cerebral vasculitis which caused extensive damage to the brain resulting in severe and permanent disability.

The issues at the hearing were narrowed because, shortly before its commencement, breach of duty was admitted on behalf of two of the general practitioners, Dr Graveson and Dr Anderson, who were the second and third defendants (“the defendants”). The concession and its effect were stated by the judge:

Cerebral vasculitis is a very rare condition involving inflammation of the blood vessels within the brain. Its aetiology is not fully understood. It is generally viewed as a mal-adapted response of the body’s immune system to some insult. The judge cited Dr J W Schmidley (Vascular Diseases of the Central Nervous System, Chapter in Neurology in Clinical Practice, Bradley et al) who stated that it manifests itself in “progressive, cumulative and multi-focal neurological dysfunction”. The neurological dysfunction is a result of reduction in blood flow to the brain cells. If it drops to a level where cellular integrity fails, cell death, which is irreversible, occurs.

The judge carefully set out the sequence of events and made findings, which are not challenged in this court, as to when the brain damage occurred. An MRI scan was carried out at Frenchay Hospital on 22 April 1998. At 14:30 hours on 27 April, the claimant underwent cerebral angiography. A further MRI scan was carried out on 6 September 1999. It was important to establish the time at which damage occurred. The judge concluded:

The judge then considered the claimant’s left-handed visual deficit and concluded:

The judge restated the issue on the basis of his findings:

The court has been referred, as was the judge, to the neurological observation charts and the clinical history sheet for 27/28 April and the course of events is substantially agreed:

At 14:00 hours, the claimant underwent an angiogram. That confirmed an “acute deterioration today”. At 16:00 hours, it was noted that the patient “has deteriorated over past couple of hours”.

Told him of D of cerebral vasculitis

Explained that we are very concerned about her condition and that she may not recover

On the point remaining in issue, evidence was given on behalf of the Claimant by Dr J Bamford, Consultant in Neurology and Cerebrovascular Medicine, and by Professor C Warlow, Professor of Medical Neurology. On the question of causation, Professor Warlow described himself as a “committed don’t know”. (The court has not been supplied with a transcript of his oral evidence.)

Against that background and on the evidence, it was and is submitted by Mr Havers QC on behalf of the defendants:

The judge stated his conclusions at paragraphs 42-48:

Secondly he accepted that the improvement that in fact took place from 2100 on 27 April could not be attributed to a dose of cyclophosphamide given at 1800 hours. He also accepted that the improvement at 18:30 was due to the mannitol given at 17:15, and that mannitol is a short acting drug used to reduce intercranial pressure in an emergency that would have been effective for two hours at the most. In those circumstances it must follow that the improvement on the 27 April was spontaneous.

Thus I am bound to conclude, contrary to the argument developed by Dr Bamford in the course of his evidence that her condition was of such a nature that spontaneous recovery was unlikely, that the Claimant’s condition was capable of spontaneous improvement. Secondly there is no evidence to suggest that the administration of cyclophosphamide had a beneficial effect, given that her condition had stabilised by 24:00 hours on the 27 April. It follows the course that the disease ran does not provide support for the contention that there would have been a more favourable outcome, and in particular that the deterioration on the 27 April would have been avoided, had cyclophosphamide been administered on 24 April.

Accordingly there remains only the assertion by Dr Bamford based upon his own clinical experience and that of others, that earlier administration of the drug regime would have made a difference. But as mentioned above his experience was limited to a single case in which as he put it, no doubt choosing his words with care,

His discussion with colleagues was with those who treat systemic vasculitis rather than the form isolated within the central nervous system. Finally, and as was submitted on behalf of the defendants, the literature is of very limited assistance given the very small numbers involved and the absence of therapeutic trials.

What if any reliance can be placed upon the fact that the treatment used in this case is accepted by neurologists as the appropriate treatment for cases of cerebral vasculitis ? The position was succinctly summarised by Professor Warlow who explained that the combination of steroids followed by cyclophosphamide is the approach that he would use “In the absence of any thing else to do. And on the basis of fragmentary evidence.” Whilst it is obviously a rational approach based as it is upon the approach adopted by those treating systemic vasculitis, I do not consider that it provides a basis upon which to conclude that it would probably have been effective in the Claimant’s case.

Given the clear evidence that administration of steroids did not prevent the deterioration on 27 April, and secondly that the improvement following the deterioration cannot be attributed to cyclophosphamide and that there is no other evidence of cyclophosphamide having had an effect, I am driven to the conclusion that the Claimant has failed to prove that to have commenced treatment at an earlier stage would have affected either the course that the disease ran or its outcome. It follows that her claim must fail”.

The issue is as to what would have been the course of events if the drugs had been administered earlier than they were: two days in the case of steroids and three days in the case of cyclophosphamide. On a balance of probabilities, would the substantial deterioration in the claimant’s condition which occurred on 27 April have occurred ? Mr Havers emphasises the need to look at the facts of the particular case. The judge was correct to reach conclusions on those facts in paragraphs 43 and 44 of his judgment, he submits, before, as he did in paragraphs 45 to 47, considering the broader picture.

In my judgment, any consideration of the effects of taking a drug upon an illness or disease, and of when those effects are likely to occur, requires knowledge of the illness and of the drug. In many cases the characteristics of the illness, and of the drug, and of the effect of one upon the other, will be very well known and the effect of administering the drug, and when the drug will have that effect, can be predicted with near certainty or a high degree of probability. In this case, the drug is well known; it is used to treat cancer and also it is toxic and may have adverse side effects. What is far less well known is the disease or condition cerebral vasculitis which fortunately is very rare. Dr Bamford had only treated one case similar to this one and Professor Warlow none at all. The effect of cyclophosphamide on the condition, and when that effect will occur, cannot be predicted with certainty, as Dr Bamford acknowledged in evidence. The facts of the particular case can only be considered, however, in the context of medical knowledge and opinion on those matters, whether that knowledge and opinion are complete or only partial and provisional.

The issue was whether, on a balance of probabilities, the achievement and maintenance of stability in the claimant’s condition after the serious deterioration on 27 April was spontaneous or was caused by the steroids and cyclophosphamide, either alone or in combination. Professor Warlow was no doubt mindful of these considerations when he commenced his report of 7 August 2003 by stating that: “because isolated cerebral vasculitis is such a rare disorder, to prepare this report I have had to rely less on experience of a handful of cases in my professional lifetime and more on the original scientific literature”. It does not necessarily follow that the judge’s conclusion was wrong, but he was not in my judgment correct to relegate his consideration of general medical knowledge until after he had expressed the conclusion that the improvement, or stabilisation as it was more appropriately described by the judge at paragraph 44, was spontaneous.

The medical literature produced to the judge was remarkably consistent in its view of the disease and of the effect of steroids and cyclophosphamide. Dr Bamford referred to it on several occasions when giving his evidence and plainly had it in mind when expressing his opinions. Dr J W Schmidley, already mentioned, stated in the introduction to his chapter:

In a chapter entitled “Neurology of the Inflammatory Connective Tissue Diseases” (from Major Problems in Neurology (Saunders 1999)), specifically mentioned by Professor Warlow in his report, Frans G I Jennekens and Louis Cater stated under the heading “Therapy and Outcome”:

In 1990, Dr Graeme J Hankey wrote an article entitled “Isolated Angiitis/Angiopathy of the Central Nervous System” (Cerebrovascular Disease, 1991, Volume 1). In the Abstract he stated:

Medical opinion is thus consistent in the view that, in its serious form, the illness is fatal without the intervention of drugs, that steroids and cyclophosphamide are the appropriate method of treatment and that they often achieve success. That this was a serious form of the disease was, on the evidence as it emerged, in my judgment the likely conclusion, had the judge addressed the question. It was the clear opinion of Dr Bamford, stated more than once in his evidence. It is confirmed by the seriousness with which those treating the claimant regarded the situation at 17:00 hrs on 27 April. Mr Havers accepted that the judge had to treat the case as a very serious example of the disease. It would be difficult to argue otherwise having regard to the medical literature and the disability, by stages, sustained.

Against this background, the judge should not have dismissed the literature in the summary way he did. In considering whether the achievement of stability was spontaneous or was caused by the drug therapy, the medical opinions emerging from the literature submitted to the judge require serious consideration. In the context of an illness which has fatal consequences in the absence of the administration of drugs, the maintenance of stability after the combination of steroids and cyclophosphamide has been administered cannot be ignored.

I turn to the judge’s reasoning at paragraphs 42 to 44. For the claimant, Mr Grace QC submits that the judge has in important respects misconstrued the evidence of Dr Bamford. Indeed, that is the main strand of his submission. Without considering the alleged misconstruction, however, the reasoning in those paragraphs cannot in my judgment stand. The “improvement” on 27 April mentioned by the judge is, on the evidence, better described as a stabilisation, a word used by the judge in paragraph 44. A serious deterioration in the claimant’s condition occurred on 27 April but, as the judge stated, “her condition had stabilised by 24:00 hours”. It does not, with respect, “follow” from the stabilisation achieved by that time that “the course the disease ran does not provide support for the contention that there would have been a more favourable outcome, and in particular the deterioration on 27 April would have been avoided, had cyclophosphamide been administered on 24 April”. That ignores the achievement, in the case of a normally fatal illness, of the maintenance of life and of stability once cyclophosphamide in combination with steroids had been administered. The judge has not mentioned, or it appears, taken into account the views consistently expressed in the medical literature, that, without treatment, the serious form of cerebral vasculitis is normally fatal in its consequences.

On the basis of the medical evidence and literature, it is entirely possible, to put it no higher, that the stability was achieved by the administration of cyclophosphamide with or without steroids. The fact that the drug could not have taken effect before midnight on 27 April does not mean that it could not have been effective after that in preventing further deterioration. It follows that the possibility that its administration on 24 April would have prevented the serious deterioration on 27 April cannot be ignored. The underlying neurological condition had been stable, as found by the judge, from 22 April until the serious deterioration on 27 April, and the presence of stability, even if without cyclophosphamide, at midnight on 27 April, does not complete the picture. The judge appears to have assumed that the stability achieved by that time was permanent. That cannot be assumed and the subsequent course of events must be taken into consideration when deciding whether the drugs had a beneficial effect and would have had a beneficial effect if administered earlier.

Dr Bamford expressed his opinion on the central point at issue in answer to the question: “If she had not received either steroids or cyclophosphamide, what do you think the outcome would have been ?” Dr Bamford stated: “I believe she would have continued to accrue neurological deficit and probably died, as happened in the pre-treatment era”. (Page 177).

When cross-examined as to what in his opinion had halted the disease process, Dr Bamford stated: “It’s the use of cyclophosphamide after the steroids. The treatment process is the combination”. He had stated earlier (p175) that the acute anti-inflammatory action was going to come from the steroids and that cyclophosphamide was much more about the suppression of the underlying disease. Asked whether there was evidence which pointed to the administration of cyclophosphamide as having been effective, Dr Bamford stated (p188):

Dr Bamford was asked in detail about the comparable case he had treated. He stated (page 192):

When, in re-examination, he was asked when the “favourable response” to the therapy commenced he stated: “I believe it commenced around late on 27 or early 28 as I said following which there was both stability and lack of accrual of further deficit”. Asked whether the improvement was because of the therapy or spontaneous, Dr Bamford stated (page 219):

In that evidence, Dr Bamford appears to me to have been expressing an opinion which was relevant, coherent and cogent.

As to the time when the drugs are to likely to have taken effect after their administration, Dr Bamford put it, in the case of steroids as 3 to 5 days, and this was not challenged in cross-examination. As to the cyclophosphamide, he stated in his report, paragraph 7.6.1:

In his evidence, Dr Bamford referred to the other patient he had “treated personally with this degree of aggressive presumed vasculitis”. He said that “the treatment appeared to have worked within 48 hours” but added “which bit of the treatment [steroids or cyclophosphamide], I don’t know”. He repeated (page 217) that it was “the treatment as a whole” which had to be considered. He agreed that the response on 27 April was not likely to have been due to cyclophosphamide.

Mr Havers relied (as did the judge) on a passage in cross-examination (pages 184-186):

I am sorry. I beg your pardon.

I suppose, or whether --- or whether it did. I don’t know. (Pause) I can only observe what treatment she was given, and the apparent response to it. I don’t think there is anything there that allows me to disentangle the two.

No. I’m sorry. I disagree with that. I mean, she’d been given five days of … (indistinguishable)… prednisolone by that point.

I see.

What I’m saying – and I think you must just disentangle improvements on the evening of 27^th from stabilisation, because they’re two different things.

If we look over the page to 298, we know, because we can see from the chart on 298 that as from 2400 hours on 27^th her condition indeed stabilised.

(Pause) Yes, I agree with that, that the stabilisation began around that time – in the sense that there is no further fluctuation and no accrual of further deficit.

And she remained stabilised thereafter.

And she remained stabilised thereafter.

(Pause) In fact, a patient’s condition when a patient has cerebral vasculitis, can improve spontaneously without any treatment at all, can it not?

If you are talking about cerebral vasculitis in toto, yes.”

Mr Havers relies in particular on two answers. First, it is claimed that when Dr Bamford stated that he did not think there was anything there that allowed him to disentangle “the two”, he was distinguishing between the treatment given and the apparent response to it. The second, when asked about spontaneous improvement, his answer, the last in the passage quoted, was: “If you are talking about cerebral vasculitis in toto, yes”.

As to the first point, I see great force in the submission of Mr Grace, that by his reference to “disentangling the two”, Dr Bamford was referring to the two drugs, the steroids and cyclophosphamide. His earlier long answer had referred to the combination of treatments, a point made elsewhere in his evidence, and referred to not being able to tell exactly when cyclophosphamide “kicked in”. Just before the statement relied upon, he had again referred, in an answer which was interrupted, that he did not know at what point cyclophosphamide kicked in. The reference to “treatment” and “response” appears to me to have been in effect in parenthesis and it was his difficulty in disentangling between the effect of the two drugs that Dr Bamford was repeating. I would have understood the answer in that way. Dr Bamford was not asked further questions to clarify the point.

As to the second point, Dr Bamford’s use of the expression “in toto” in my view deprives the answer of the sting relied on. Not surprisingly, he had been asked about the two forms the disease may take and it was not in issue that the benign form can improve spontaneously.

Of less importance, given the central issue, is the judge’s reliance on Dr Bamford’s answer to the question whether the steroids would have prevented the deterioration had they been given two days earlier that “it is unlikely because you are simply extrapolating from what happened”. The sequence of questioning was:

Following a meeting between the neurological experts Dr Bamford and Professor Warlow on 1 October 2003, a series of answers had been recorded, incorporating, where there was disagreement, the opinion of Dr Bamford.

Dr Bamford is of the opinion that if treatment of methylprednisolone was started on 20 April 1998 and cyclophosphamide on 24 April 1998 then on the balance of probability the deterioration on the 27 April 1998 would not have occurred. The additional neurological damage accruing on 27 April 1998 would have been prevented and she would therefore have had a better outcome. The reason for his answer is that a review of the charts from the 27 April 1998 onwards shows that, following treatment, Mrs Roughton did not accrue any further focal neurological deficits and that her condition appears to have stabilised within 24 hours of the treatment with cyclophosphamide. He believes it was therefore probable that the same pattern would have occurred if that treatment (ie with methylprednisalone and cyclophosphamide) had been given as set out in the question.

Both experts have discussed this with their colleagues. Professor Warlow’s opinion is that he would expect to see an improvement in a patient’s condition within a week or probably more. Dr Bamford’s opinion is that, he would say the speed of response to cyclophosphamide is variable. He agrees that in sub-acute or chronic cases the response maybe over days or longer but his opinion, based on small personal experience and discussion with colleagues who deal with systemic vasculitis, is that in very aggressive cases a response may be seen within 1 to 2 days.”

The word “improve” used in questions 6 and 10 is a misnomer in the sense that the court was concerned with the achievement of stability and the absence of further neurological deterioration. It is the claimed response to the drug therapy which is the important factor. Dr Bamford’s repeated use of word “response” in answer to question 10 is consistent with his oral evidence, though he did accept the word “improvement” in reply to a question which included that word, during re-examination (p217). The judge correctly used the word “stabilised” in paragraph 44 of his judgment and the expression “the course that the disease ran or its outcome” in paragraph 48.

Analysis of Dr Bamford’s evidence in my judgment reinforces the view that, with respect, it does not “follow” from the course the disease ran that there would not have been a more favourable outcome had cyclophosphamide been administered on 24 April. The evidence did not drive the judge to reach, as he thought it did, the conclusion he reached. The judge has not had regard to the achievement, after 27 April, of stability in a case of cerebral vasculitis where, on the evidence at it emerged at the trial, the outcome is normally fatal.

The judge did not make a clear finding as to whether this was a case of primary as distinct from benign vasculitis. The evidence that it was the more aggressive form appears to me, for reasons given earlier, to have been very strong. A conclusion that, because there was no deterioration after 27 April, necessarily, the therapy did not, and would not earlier, have worked, is not in my judgment tenable on the evidence. It could not be assumed that the stability achieved at midnight on 27 April was permanent. The stability between 22 April and early on 27 April did not prove permanent.

The question is whether, on the balance of probabilities, the earlier drug therapy postulated would have prevented the deterioration on 27 April. Whether the permanent stability achieved was spontaneous is obviously relevant to that question, it being agreed that stability as a result of drug therapy or stability achieved spontaneously are the only possibilities. The first question for this court is, however, whether it is able to decide that issue or whether it is necessary to remit the issue of causation for re-trial. Mr Grace submits that the court is an as good a position to assess the evidence as was the trial judge.

I have reluctantly come to the conclusion that remission is necessary:

The judge has failed to consider the significance of the maintenance of stability after 27 April and has wrongly confined his focus to the events of the last few hours of that day.

The evidence of Dr Bamford is an important part of the claimant’s case and its assessment in the context of a trial, and with the correct issues in mind, is necessary.

Notwithstanding the detailed reference, and the sustained comment, I have thought it necessary to make upon the evidence and the judge’s approach to it, I am not able to hold that it is appropriate that this court should take upon itself the task of resolving the issue of causation. On that issue, and that issue alone, a reassessment of the evidence as a whole is required. The judge’s findings on other issues, which have not been the subject of challenge, will stand.

Evidence may be called. I am conscious that this gives the defendants a second opportunity to put forward a positive case which they could have, but did not, put forward at the trial on the issue the judge had identified. This could work hardship for the claimant but that possibility does not overcome the need for reassessment by a fact-finding tribunal.

To the extent indicated, I would allow this appeal

I agree that the issue of causation must be remitted for retrial, for the reasons which Pill LJ has given. I also agree with the observations of Hooper LJ (whose judgment I have had the benefit of reading in draft).

I agree with Pill LJ that the issue of causation should be remitted for a further hearing. The question which the judge had to resolve was whether the claimant had shown on the balance of probabilities that the substantial deterioration in the claimant’s condition which occurred on 27 April would not have occurred if the drugs had been administered earlier than they were: two days in the case of steroids and three days in the case of the cyclophosphamide. To resolve that issue, the judge had to decide whether, on the balance of probabilities, the drugs are generally effective to “treat” the form of vasculitis from which the claimant was suffering and if so in what period of time.

In deciding this issue the events of 27 April and thereafter were relevant (although not necessarily decisive). If the undisputed absence of any further deterioration after that date was due to the administration of cyclophosphamide on that day (alone or coupled with the administration of steroids), then that showed the effectiveness of the drugs. If the vasculitis was, on the balance of probabilities, of the kind which would cause death in the absence of the administration of the drugs, as Dr Bamford opined, then the absence of any further deterioration would have been caused by the drugs and would probably have been caused, so he thought, within less than 48 hours. Unfortunately the judge concentrated, as Pill LJ has explained, on the improvement by about midnight and not on the absence of any further deterioration. The improvement by midnight led the judge to conclude that the claimant’s condition was capable of spontaneous improvement and that the administration of cyclophosphamide on that day was not therefore shown to have had a beneficial effect. If he had concentrated on the absence of any further deterioration, his conclusions may well have been different.