Wardlaw v Farrar

This is an appeal by the claimant Peter Wardlaw against the order of Judge Singer in the Manchester County Court on 14th August 2002, when he awarded him only £1,000 damages in respect of a claim for damages for clinical negligence arising out of the death of his wife.

Mr Wardlaw sued as widower and personal representative of his wife in a claim brought under the Fatal Accidents Act 1976 and the Law Reform (Miscellaneous Provisions) Act 1934. She died on 26th September 1997 in the Manchester Royal Infirmary. Mr Wardlaw initially issued proceedings against both the relevant National Health Trust and the present respondent, Dr Farrar, who was her general practitioner. However, he discontinued the proceedings against the Trust because he was advised that causation would be very difficult to prove.

Mrs Wardlaw was 69 when she died. She was a retired hospital domestic and a known hypertensive for many years. She was an ex-smoker, with a long history of anxiety and depression. Her history of multiple aches and pains and chronic headaches stretched back to the early 1970s.

Between mid-July and mid-September 1997 she saw GPs from her registered practice on a number of occasions. The crucial date was 14^th September, when the judge found that Dr Farrar had been negligent in the examination which he conducted that day. He did not consider a diagnosis of pulmonary embolism (“PE”), which eventually resulted in her death, and he did not arrange her admission to hospital for further investigation. In the result her admission to hospital was delayed until she was seen again by a locum doctor on 21st September. She was then admitted to hospital in something of an emergency. In the intervening period she remained in pain and suffered from breathlessness. She spent most of that week in bed.

The judge found that Dr Farrar’s negligence was responsible for a week's pain and suffering during the delay in the admission to hospital, but that that delay did not increase the risk of her death. It did not therefore materially contribute to her death, which occurred some days after her admission.

She was admitted to hospital at 11.30am and treated with anti-coagulation therapy for a suspected diagnosis of pulmonary embolism. Heparin treatment started at 6.15pm that day. At 11.10pm she sustained a sudden collapse, which was associated with increased pain and increased shortness of breath. The following day (22^nd September) a lung ventilation and perfusion scan was performed. This confirmed the presence of numerous small pulmonary emboli, but no large emboli were evident. Warfarin treatment was started that day. Over the next 48 hours her general condition improved. The anti-coagulation therapy was working, and it was suggested that she might go home soon.

On 25^th September, following a ward round, she collapsed suddenly with worsening abdominal pain. She was pale and clammy, with a fast heart rate and low blood pressure. A Doppler ultrasound scan of the veins in her leg confirmed the presence of a thrombus in her right popliteal vein. The following day (26^th September) a pulmonary angiogram was performed. This revealed massive bilateral pulmonary emboli, with near total occlusion of the right pulmonary artery and excessive emboli on the left. Despite a mechanical embolectomy she deteriorated suddenly and died at 6pm that evening. A doctor who reviewed her at 4pm was of the view that a thrombus in her right leg was the source of the multiple pulmonary emboli.

There were two experts in the case. Dr David Evans, a consultant physician, was called by the claimant, and Dr David Hill, a consultant cardiologist, was called by the defendant. In their main reports they concentrated principally on the question whether there was any negligence on the part of Dr Farrar or of those treating Mrs Wardlaw when she was in hospital. Dr Evans concluded that there was, and Dr Hill concluded that there was not.

On the crucial question of causation in Dr Farrar’s case, their differing views were most succinctly put in their joint report in answer to questions from the parties’ legal representatives, when they dealt with the delay after 14th September 1997:

Dr Evans relied strongly on an article published in The Lancet in April 1999 which summarised the findings of a newly established International Co-operative Pulmonary Embolism Registry (“ICOPER”). The organisers of this venture registered 2,454 consecutive eligible patients with pulmonary embolisms in 52 hospitals between January 1995 and November 1996. Hospitals in five different European countries (Italy, France, Switzerland, Poland and Belgium) registered their data at a co-ordinating centre in Bologna, while hospitals in two North American countries (USA and Mexico) registered their data in Boston. Most of the results were recorded from hospitals in Italy and the United States. Four of the countries participating in the scheme registered about 400 outcomes between them.

The reason for this exercise were that despite advances in diagnosis and therapy, pulmonary embolism (“PE”) remained poorly understood ten years ago, with wide discrepancies in reported rates of mortality and recurrence.

The finding of the study that attracted most attention at the trial was that out of 96 patients who were haemodynamically unstable at the time of presentation 56 (58.3%) were dead within three months. In contrast, the mortality risk in relation to the 2,093 who were haemodynamically stable on presentation was only 15.1% (representing 317 fatalities within the same timespan).

The problem the experts encountered at the trial was that the study furnished no definition of the expression “haemodynamically unstable”. Nor did the guidelines published by the British Thoracic Society. Dr Evans agreed that the words might be properly used to describe a patient who appeared ill on presentation, with low blood pressure, breathlessness and a quickened heart beat, who might possibly be fainting at the time. Dr Hill said that what the phrase meant in real life was the perception that the blood clots had disturbed the blood circulation and the power of the heart to beat because they were obstructing the blood vessels in the lungs. Since pulmonary circulation represented half the circulation of blood through the body, it came as no surprise to him that there was a high death rate among the members of the group who were haemodynamically unstable. The difficulty lay in selecting the patients to put into this group in the first place, and the ICOPER study did not help with this. Dr Hill suspected that the phrase reflected how the patients appeared to the clinicians in the different places where they were examined, so that in the absence of agreed criteria for haemodynamic instability it was inappropriate to draw any particularly firm conclusions from this statistic.

One table in the study showed variables that were both significant and independently predictive of an increased three-month mortality rate. These, with what was called the hazard ratio of each, were

They agreed that when Mrs Wardlaw presented to Dr Farrar on 14^th September she was haemodynamically stable. They did not agree about her status on admission to hospital on 21^st September. Dr Evans accepted that her blood pressure was stable at 139/85 when she was first admitted that day, although her heart rate (130 beats per minute) was significantly raised. He thought that the blood pressure reading of 100/60 at 11pm that night, associated as it was with an episode of acute breathlessness, was indicative of haemodynamic instability. He accepted that on 21^st September the jugular venous pressure was not elevated, and that this was a feature one would expect in a patient in a severe state. He also accepted that there was no right ventricular heave on admission, and its absence would militate against the presence of a large PE. The two ECG results on admission showed abnormal tachycardia, but there was no evidence of marked acute heart strain, such as one would expect if a sizable PE was present. There was nothing unusual about Mrs Wardlaw’s temperature that day. On the other hand she had a very low blood oxygen (PO2 = 50). All in all, Dr Evans would not say that she was haemodynamically stable on admission with a heart rate of 130 and a PO2 of 50. She was obviously ill, but he was not sure if she fulfilled a definition of instability. She suffered from low blood pressure later in the evening.

Dr Hill classified Mrs Wardlaw as “borderline” haemodynamically unstable on admission. She did not look like someone in dire distress, but PE was a distinct possibility. If a patient was in dire distress caused by a PE, he would expect her to be breathless, with a reduced blood pressure and tachycardia, and to look ill without necessarily having chest pain. The veins in her neck would be raised from the back pressure of the obstruction from the heart. The simple ECG tests would be likely to be very abnormal, and this did not happen in Mrs Wardlaw’s case (not even when she was very ill at the end of her life).

The judge accepted Dr Hill’s evidence that the degree of a person’s instability was of great importance. Because the doctors who contributed to the ICOPER study were not working from an agreed definition of instability, he found that the results of the ICOPER study could not be used in any given case to predict or state the likely outcome for a particular individual of treatment at any given time. This appears to me to be a conclusion he was entitled to reach on the evidence before him.

Another important part of the medical evidence related to the conclusions that might be drawn from the fact that both heparin and warfarin had been prescribed for Mrs Wardlaw after she was admitted to hospital on 21^st September and that from 24^th September onwards her blood was fully anti-coagulated. On being reminded of the readings on 24^th and 25^th September Dr Evans accepted that these two drugs were now doing their job effectively. In those circumstances there should not have been any new thrombus formation, but he believed that a pre-existing unstable thrombus might still have produced a fatal PE. Mrs Wardlaw had three episodes of collapse during her in-patient stay, and a large embolus could have broken off an unstable thrombus at one of these times.

He said that when heparin was prescribed, it might take a week before a thrombus would stabilise. The role of heparin and warfarin was to prevent new thrombi forming, and the readings did not show whether there was an unstable thrombus already in the bloodstream.

Dr Hill agreed that heparin and warfarin did not dissolve clots: they prevented larger clots being formed. Logically there must have been a persisting thrombus throwing off an embolus after she was treated. In general one would expect clotting to stop within 24 hours of the start of heparin treatment, and within two days in the case of warfarin. The textbooks stated that it might take 3-5 days with warfarin, but in general medical practice it was very unusual to have a patient on warfarin for more than two days flying off clots unless there was some reason behind it. He accepted that in Mrs Wardlaw’s case there was no evidence that she had any underlying malignancy such as might in itself have made her resistant to anti-coagulation.

In this context he referred to a passage in the Oxford Textbook of Medicines which states:

Unfortunately, the common form order in clinical negligence cases to the effect that lists of learned articles or textbook passages should be exchanged between the parties was not made by the district judges at Manchester County Court when they gave relevant case management directions on two occasions. This omission led to serious difficulties. Dr Evans had referred to the ICOPER article in his initial report, but it does not appear to have been copied for Dr Hill or otherwise exhibited in the litigation until the first day of the trial. Dr Hill for his part did not mention any learned references in his report, but copies of relevant extracts from the Oxford Textbook were sent to the claimant’s advisers just before the trial started. Unhappily two pages were not copied, but nobody on the claimant’s side appears to have noticed this omission.

The standard form of order made by the High Court masters in the Queen’s Bench Division in clinical negligence cases now contains these provisions:

Now that so many relatively heavy clinical negligence actions in the multi-track are being conducted in county courts, in addition to district registries of the High Court, it is essential that best practice should be followed throughout the country in relation to case management directions in the multi-track in this specialist field. If these directions had been made in the present case, the medical literature would have been handled in a more orderly manner.

At the outset of the appeal we refused an application by Ms Ruck, who appeared for the claimant, for permission to produce the two missing pages from the text book as fresh evidence on the appeal. She wished to adduce in evidence a statement in a different article to the effect that of 300,000 – 600,000 patients who suffered a PE in the United States every year, 3 - 8% die despite treatment, either from recurrent embolism or more often with embolism complicating an underlying fatal disorder. We refused this application not only because with reasonable diligence the claimant’s advisers could have adduced this evidence from a standard text book before the trial (indeed they could have called for the two pages missing from the material sent to them by the defendant’s solicitors if they had troubled to read it), but more importantly because it appeared to add nothing to the material already placed by Dr Hill before the court. The judge knew that the number of patients who do not respond to treatment were variously described as “a very few” and “a very small number”. He also knew that part of this phenomenon could be ascribed to an established malignancy and part to some other unidentified problem. In particular he heard Dr Hill say:

Dr Evans’s approach to this issue was that he did not see why the court had to assume that Mrs Wardlaw had a chronic untreatable thrombo-embolic syndrome. He felt that she had a single unstable thrombus in her leg and another that latterly developed in her right pulmonary circulation in the form of a massive PE. He assumed that thrombosis with emboli was the cause of her illness over the last 8 – 12 days of her life. This history was too short to place her in Dr Hill’s group of rare patients (whose existence he accepted).

On the appeal to this court Ms Ruck endeavoured to extract from the transcripts an entirely new argument which had never been put to Dr Hill when he was in the witness-box nor advanced by her predecessor to the trial judge. She picked on the following questions and answers in support of her submissions:

A little later Dr Hill said that the proposition that a thrombus would grow without treatment was a reasonable assumption to make, but he also thought that there were other factors that might make this difficult to assess. That answer was not properly explored, but this passage of the evidence proceeded with Dr Hill sticking to his opinion, which the judge ultimately accepted, that in practical terms the seven day delay did not make much difference because with this particular patient anti-coagulation therapy, once it had achieved its purpose, did not have its normal effect, as the sudden deterioration on 25^th-26^th September showed.

The judge observed (at paras 23-24 of his judgment) that Dr Evans had said in a pre-trial letter that the advantage of starting to administer heparin on 14^th September would have been more to do with what he described as “the earlier intervention in the natural history of the disorder”, because he accepted that Mrs Wardlaw in fact responded unsatisfactorily to heparin treatment when it was administered at a later point. The judge commented that Dr Evans did not, however, explain what he meant by this until he gave evidence at the trial, when he said that he was relying on the statistics revealed in the ICOPER report.

This court is always willing to entertain new arguments on law on an appeal if they would not have affected the factual basis in which evidence was given at the trial. But it cannot entertain an entirely new argument based on the transcripts, when the factual evidence might have been significantly different if the point had been taken clearly in the court below. In the court below the claimant’s advisers were relying on the ICOPER statistics which the judge was fully entitled to reject as a safe way of proceeding in this particular case (see para 17 above). They were also relying on a contention that somehow or other a longer period of treatment with heparin and warfarin would have made a difference, but that, as the judge observed, was not Dr Evans’s view.

At the trial counsel who was then instructed on behalf of the claimant argued that if Mrs Wardlaw had been admitted to hospital on 14^th September, this would have allowed the anti-coagulant therapy more time to work. He did not advance any argument – nor was the point clearly put to Dr Hill in cross-examination – to the effect that a thrombus which was present on 14^th September had grown larger and more potentially harmful by 21^st September. The judge’s conclusion on this issue was expressed in these terms (at paras 29-33 of the transcript):

It will be evident from this passage that the judge was much more impressed by Dr Hill’s evidence than he was by Dr Evans’s evidence. In particular he felt that Dr Hill was drawing convincingly on his long clinical experience while Dr Evans was prone to express conclusions without supporting reasons other than his reliance on the ICOPER report. A little earlier in his judgment (at para 26) he had been critical of a particular passage in Dr Evans’s evidence in which he had based his belief that Mrs Wardlaw was haemodynamically stable that day on the supposition that Dr Farrar had examined her properly that day, a conclusion which the judge rejected. There is clear authority that this court should be very slow to interfere with a trial judge’s views on the quality of the evidence of expert witnesses whom he has had the advantage of seeing and hearing (see, for example, Wilsher v Essex Area Health Authority [1988] AC 1074, 1091). I can see no reason why we should interfere with the judge’s assessment.

Ms Ruck then endeavoured to found an argument on the decision of the House of Lords in Hotson v Berkshire Area Health Authority [1987] AC 750. She maintained that the appropriate approach of the court was to consider the probabilities as at the time “the negligent treatment started”. Given that Mrs Wardlaw was haemodynamically stable at that time, she said that the overall probabilities, based on the ICOPER statistics, were that she had an 85% chance of survival.

Her chances of survival, Ms Ruck said, were greatly reduced by the delay in treatment. She accepted that if there was some other known fact (such as a malignancy discovered post mortem), then a prognosis based on general probabilities was bound to founder. But she was unwilling to accept that the fact that Mrs Wardlaw in due course proved resistant to the usual beneficent effect of anti-coagulation therapy could appropriately be taken into account when considering the probabilities as at 14^th September. She asked us to take into account in this context her submission that it was the negligence of Dr Farrar that allowed the thrombus to grow.

I do not understand this. When the judge had to consider, on the balance of probabilities, whether Dr Farrar’s negligence (and the consequent delay in her admission to hospital) was causative of Mrs Wardlaw’s death the judge had to take into account all the relevant evidence, and the rival cases that were being put forward at the trial in relation to this evidence. The failure of anti-coagulant therapy, when it was tried, to prevent the formation of a massive pulmonary embolism (which had not been present on 22^nd September) was inevitably a material piece of evidence. While judges are of course entitled to place such weight on statistical evidence as is appropriate, they must not blind themselves to the effect of other evidence which might put a particular patient in a particular category, regardless of the general probabilities.

For these reasons I would dismiss this appeal.

I agree.

I also agree.