Saffron Paul v The Paul Wolverhampton NHS Trust

These three cases raise the question of the circumstances in which a defendant to a clinical negligence claim can be held liable for psychiatric injury (or what used to be called nervous shock) caused to a close relative of the primary victim of that negligence. The basic facts in each of the cases are that the defendant is alleged to have failed to diagnose the primary victim’s life-threatening condition. Some time after that negligent omission, the primary victim suffered a traumatic death. In two of the cases (Paul and Polmear), the shocking death occurred in the presence of the close relatives, causing them psychiatric injury. In the case of Purchase, the close relative came upon the primary victim immediately after her death, again causing her (the mother in that case) psychiatric injury. The question in each case was whether the necessary legal proximity existed between the defendant and the close relative (often referred to as the secondary victim).

This question of legal proximity was dealt with authoritatively in Alcock v. Chief Constable of the South Yorkshire Police [1992] 1 AC 310 (Alcock), where Lord Oliver described the common features of the reported cases at that time at page 411F-H. It was from the following five elements that the essential requirement of proximity had to be deduced:

Lord Oliver made clear at page 412A (and this is important here for a reason I will come to in due course) that liability of this kind can arise even where there is no “primary victim”: for example: “a parent may suffer injury, whether physical or psychiatric, as a result of witnessing a negligent act which places his or her child in extreme jeopardy but from which, in the event, the child escapes unharmed”.

The first case before us, Paul, is a second appeal, where Mr Justice Chamberlain held that Master Cook had been wrong to conclude that the claims were bound to fail. The necessary proximity to Mr Paul’s children could be established on the basis that Mr Paul’s collapse and death, which was the shocking event (I shall generally use the term “horrific event”) the children had witnessed, was the first manifest damage caused by the defendant’s negligent failure to diagnose his heart condition. The question, according to Chamberlain J, was whether Mr Paul’s death was capable of constituting a relevant event. He held that it was. Chamberlain J distinguished the two cases that Master Cook had relied upon: Taylor v. Somerset Health Authority [1993] PIQR 262 (Auld J) (Somerset) and Crystal Taylor v. A. Novo (UK) Ltd [2013] EWCA Civ 194 (Court of Appeal) (Novo). Somerset was a clinical negligence case with facts similar to Paul, but Novo was an accident case, where the claimant had been shocked by the traumatic death of the primary victim three weeks after her injury in the accident.

The second case before us, Polmear, was decided by Master Cook after he had been reversed by Chamberlain J in Paul. Master Cook said that the parties were agreed that the question on the strike out was whether the claimant parents had a reasonably arguable case that the relevant event required to satisfy the control mechanism of proximity was the collapse and death of the primary child victim, Esmee (5 or 7 months after the negligent failure to diagnose). He followed Chamberlain J and held that it was possible to identify a qualifying horrific event and that that horrific event did not have to coincide with or immediately precede the first actionable damage to the primary victim.

The third case before us, Purchase, was decided by District Judge Lumb in favour of the defendant a month before Chamberlain J’s decision. DJ Lumb held that he was bound by Novo, which meant that the claim was doomed to fail.

In each of the three appeals, Mr Charles Bagot QC and Ms Charlotte Jones, counsel for the defendants, accepted that claims by secondary victims for psychiatric injury are available in clinical negligence cases. They submit, however, that the deaths in each of these cases were separated in space and time from the negligence that occurred in a hospital or primary care setting. They cannot, therefore, be said to be “the relevant event for deciding the proximity required to establish liability under the established control mechanisms”. In essence, they submit that this court is bound by its own decision in Novo, which decided that a secondary victim cannot claim in respect of psychiatric injury sustained by witnessing any horrific event once actionable damage has already been sustained by the primary victim on an earlier occasion. Moreover, Novo approved Somerset whose facts were on all fours with these cases.

Mr Robert Weir QC and Ms Laura Johnson, counsel for the Paul claimants, submitted that Chamberlain J was right to distinguish Somerset and Novo. They argued that the relevant event or trigger for the liability to the secondary victim had to be a single event “that was the damage that it was the duty of the defendant to protect the primary victim against when the damage first becomes manifest or evident”.

Mr Henry Pitchers QC and Mr Oliver May, counsel for the Polmear claimants, and Mr David Tyack QC and Mrs Esther Gamble, counsel for the Purchase claimant, took a different stance as to the relevant event. They submitted that any horrific event caused by a breach of duty to the primary victim was sufficient to give rise to legal proximity and liability to a secondary victim satisfying Lord Oliver’s five elements, whether or not damage to the primary victim had occurred or manifested itself at an earlier time.

In these circumstances, we have to decide whether a claimant, who sustains psychiatric injury as a result of witnessing the death or other horrific event suffered by a close relative as a result of earlier clinical negligence, can claim damages for that psychiatric injury. The question turns on the relevance of any time intervals between the clinical negligence, the damage caused by it, and the horrific event that ultimately causes the psychiatric injury to the claimant. The parties have put forward three possible answers. They suggest that, as a matter of law, a defendant to a claim for damages for clinical negligence can be liable to a secondary victim who has suffered psychiatric injury by witnessing the death or other horrific event affecting the primary victim and caused by the negligence: (a) only when that horrific event is the damage completing the primary victim’s cause of action in negligence, or (b) only when that horrific event is the first manifestation of damage to the primary victim caused by the negligence, or (c) whenever that horrific event occurs. The defendants say that the answer is (a). The Paul claimants say that the answer is (b). The Polmear and Purchase claimants say that the answer is (c).

The resolution of these questions necessitates a careful examination of the authorities so as to understand precisely what they decided. The difficulty is exacerbated by the fact that the decisions of the House of Lords on which all parties rely and the decision of the Court of Appeal in Novo were cases about horrific events caused by accidents rather than clinical negligence. The issue we have to decide has not been specifically considered before in this context in the Court of Appeal.

I have concluded, in brief, that the five elements required to establish legal proximity in secondary victim cases apply as much to clinical negligence cases as they do to accident cases. The question of what is a relevant horrific event is not dependent either on the completion of the primary victim’s cause of action for negligence or the first manifestation of injury to the primary victim. For a secondary victim to be sufficiently proximate to claim for psychiatric injury against the defendant whose clinical negligence caused the primary victim injury, the horrific event cannot be a separate event removed in time from the negligence. If the negligence and the horrific event are part of a continuum as they were in Walters, there is sufficient proximity. Novo is binding authority for the proposition that no claim can be brought in respect of psychiatric injury caused by a separate horrific event removed in time from the original negligence, accident or a first horrific event. I accept that, although there is no logical reason for these rules, they are the way Auld J in Somerset and the Court of Appeal in Novo built upon the five elements and adapted them to the clinical negligence context. If I were starting with a clean sheet, I can quite see why secondary victims in these cases ought to be seen to be sufficiently proximate to the defendants to be allowed to recover damages for their psychiatric injury. Since, however, this court is bound by Novo, it is for the Supreme Court to decide whether to depart from the law as stated by Lord Dyson in that case.

I shall proceed now to explain the facts of the three cases in a little more detail and to consider the relevant authorities in chronological order, before summarising briefly the reasoning of the courts below and determining the issue raised by the appeals.

Mr Paul suffered a heart attack and collapsed on 26 January 2014 when shopping with the second claimant (aged 12) and the third claimant (aged 9). His daughters saw him fall backwards and hit his head on the floor. The second claimant was so distressed that her attempts to ring her mother and an ambulance failed. A member of the public called an ambulance. The third claimant contacted her mother but could not be understood due to her distress. Both claimants saw a man holding their father’s head; there was blood on his hands. Their mother subsequently arrived. The second and third claimants heard their mother screaming their father’s name. They saw the ambulance crew put a foil blanket over him. Paramedics were performing chest compressions. The ambulance arrived at 15.57. It left the scene at 16.28. Mr Paul was declared dead at hospital at 16.51. Mr Paul’s heart attack was caused by ischemic coronary artery atherosclerosis.

Mr Paul suffered from type II diabetes and from complications of this condition. On 9 November 2012 he was admitted to New Cross Hospital in Wolverhampton complaining of chest and jaw pain which radiated to the left arm. He was treated for acute coronary symptoms and was discharged on 12 November 2012. It is the claimants’ case that the defendant was negligent in failing to perform coronary angiography in November 2012 which would have revealed the coronary artery disease and which could have been successfully treated by coronary revascularisation.

As a result of the events on 26 January 2014, the second and third claimants sustained psychiatric injury.

Esmee Polmear, aged 7, was seen by her GP on 19 August 2014 with a history of strange episodes during which she could not breathe, appeared pale and turned blue after a few minutes. On 10 September 2014, Esmee and the claimants returned to the GP because of worsening symptoms. Esmee was referred to a paediatrician at the hospital and was seen on 1 December 2014 in the presence of the claimants. As a result, from 21 to 22 January 2015 Esmee underwent ambulatory ECG monitoring, which revealed no episodes of shortness of breath. The reviewing paediatrician concluded that Esmee’s symptoms were likely to be related to exertion and were physiological “with nothing to suggest an underlying abnormality of her cardiac rhythm”. Esmee was seen again by her GP on 21 April 2015 accompanied by the second claimant. Esmee was re-referred to the paediatrician at the hospital but the referral did not take place due to her death on 1 July 2015, the cause of which was pulmonary veno-occlusive disease.

On 1 July 2015, Esmee was due to attend a school trip but did not feel well. It was agreed that the second claimant would meet Esmee at the beach to take her back to school if required. When he later went to the beach, Esmee was not present. The second claimant found Esmee with a teacher and another pupil. Esmee looked tired, pale and was breathless. Esmee wanted to sit down but was encouraged to try to walk. At one point she stopped and vomited. The first claimant joined them. The second claimant resumed the walk to the school but Esmee seemed frightened at the thought of walking and had to stop frequently, causing him to carry Esmee to the school. She was white and clammy with some blueness around her lips. At the door of the school, Esmee said that she felt faint. The second claimant reassured and comforted her. He walked away but received a call asking him to return. On doing so, he saw Esmee lying on the floor and a member of staff providing first aid. The second claimant took over and attempted to give Esmee mouth-to-mouth resuscitation. She was not breathing. The first claimant ran to the school and saw Esmee lying on the floor with members of staff attempting resuscitation which she could see was not working. Paramedics arrived and attempted resuscitation, which was witnessed by both claimants. The claimants went with Esmee in an ambulance to hospital. Attempts to revive Esmee continued at the hospital but they were unsuccessful.

As a result of witnessing the collapse, unsuccessful attempts to resuscitate and the death of Esmee, the first claimant has developed post-traumatic stress disorder and major depression. The second claimant has developed post-traumatic stress disorder, and major depression with addictive behaviour.

The defendant admits that Esmee’s condition should have been diagnosed by mid-January 2015.

Evelyn Purchase, aged 20, died on 7 April 2013. The cause of her death was extensive bilateral pneumonia with pulmonary abscesses.

On 28 January 2013, Evelyn visited her GP with acute sinusitis. In February Evelyn continued to feel unwell. She lost her appetite, resulting in weight loss. On 28 March 2013, Evelyn visited her GP and was prescribed medication for oral thrush and for a skin infection. She subsequently developed a cough and mouth ulcers. She lost her appetite and stopped eating. By 4 April 2013, Evelyn was weak and generally unwell. The claimant took Evelyn to the out-of-hours clinic where she was examined by the defendant. Evelyn had difficulty walking into the clinic as a result of weakness, dizziness and difficulty in breathing which was rapid, shallow and noisy. The diagnosis made was respiratory tract infection with pleuritic pain, oral thrush and “depressed”. Antibiotics and an antidepressant were prescribed. Evelyn was advised to contact her own GP if the problems did not resolve.

Evelyn’s condition remained the same, save by 6 April 2013 she complained of heart palpitations. That evening the claimant attended a pre-planned event in London with her younger daughter. She discussed staying at home, but Evelyn insisted she kept to her plans. Evelyn’s father remained at home with her.

The claimant returned home at 4.50am on 7 April 2013. She found Evelyn lying motionless on her bed with the house telephone in her hand, staring at the ceiling not moving. Her skin was slightly warm, she looked alive but was not moving or blinking. The claimant felt stunned, panicked and began screaming. She was joined by her younger daughter and her ex-husband. All were screaming. The claimant attempted to call 999 but the phone would not work. The younger daughter called 999 and the family were advised to give Evelyn cardiopulmonary resuscitation. In attempting mouth-to-mouth resuscitation, the claimant opened Evelyn’s mouth but this caused blood and bodily fluids to spill out of the mouth and nose. The claimant tipped Evelyn’s body to one side and more fluid spilled out. Increasingly aware that her efforts would be in vain, the claimant attempted resuscitation until the arrival of paramedics. The paramedics’ attempts at resuscitation were unsuccessful and the claimant was told that her daughter had died.

The claimant realised that she had a missed call from Evelyn on her mobile phone and a voice message. It was the sound of Evelyn’s dying breaths which continued for four minutes and 37 seconds. This caused the claimant to run out of the house and stand screaming in the street. The call was timed at 4.40am, and concluded approximately five minutes before the claimant saw Evelyn.

The claimant has developed post-traumatic stress disorder, severe chronic anxiety and depression with continuing symptomatology. It is the claimant’s case that Evelyn had severe pneumonia when seen by the defendant on 4 April 2013. It is alleged that there was a negligent failure to properly assess and treat Evelyn’s symptoms.

In McLoughlin, a claimant mother was taken to the hospital to see her severely injured husband and three children some two hours after a road accident caused by the defendant. Lord Wilberforce described the issue at page 417H-418A as being whether someone who was not present at the scene of grievous injuries to her family but who came upon those injuries at an interval of time and space, could recover damages for nervous shock. He said at page 420F that foreseeability had to be accompanied and limited by “the law’s judgment as to persons who ought, according to its standards of value or justice, to have been in contemplation”. Lord Wilberforce said expressly at page 421B that he had to bear in mind that cases of nervous shock and the possibility of claiming
damages for it, were “not necessarily confined to those arising out of accidents
in public roads”. Before explaining the “aftermath doctrine” at page 422D-E, Lord Wilberforce said in a much-quoted passage that: “[a]s regards proximity to the accident, it is obvious that this must be close in both time and space. It is, after all, the fact and consequence of the defendant’s negligence that must be proved to have caused the “nervous shock””. He then said that he accepted, by way of reinforcement of the “aftermath” cases, an analogy with rescue situations, but subject to the qualifications he had set out, “a strict test of proximity by sight or hearing should be applied by the courts”.

It is to be noted that, in hindsight, Lord Wilberforce was beginning the process of creating the five elements later stated by Lord Oliver. He was, however, conscious that he was limiting liability for psychiatric injury caused in all situations not just road traffic accidents. The limitation introduced that is most relevant to this case is that the fact and consequence of the defendant’s negligence (in that case the accident that was the horrific event that caused the nervous shock) had to be close in time and space to the moment when the claimant was caused the nervous shock (in that case, the time when the claimant saw her family in distress in the hospital). Thus, as Lord Wilberforce had indicated in stating the issue that the House was deciding, McLoughlin was really an aftermath case.

As Lord Keith explained at page 396E in Alcock, Lords Bridge and Scarman rested their finding of liability in McLoughlin on reasonable foreseeability, and Lords Edmund-Davies and Russell said nothing inconsistent with the speech of Lord Wilberforce.

Alcock concerned the psychiatric injury caused to friends and relations of those killed and injured (or in one case not injured at all) in the Hillsborough disaster. The defendant was sued for negligent policing. The claimants argued that reasonable foreseeability of the risk of psychiatric injury was all that was required to make good their claims. Lord Keith giving the first speech made clear at page 396H that a requisite relationship of proximity between the claimant and the defendant was necessary in addition to reasonable foreseeability of injury. Lord Ackner at page 402D-403H described three elements of proximity as the class of persons, the proximity of such persons to the accident – in time and space, and the means by which the shock had been caused. Lord Oliver at page 406G-H described the duty question as involving the twin questions of (a) whether injury of this sort was a reasonably foreseeable consequence of the acts or omissions constituting the breach of duty to the primary victim, and (b) whether there existed between the defendant and each claimant that degree of directness or proximity necessary to establish liability.

Lord Oliver then considered a number of cases where the claimant had been personally involved in the incident out of which the claim arose, either through the direct threat of bodily injury or in a rescue. He said that the analysis was more complex where, as in Alcock, the injury was attributable to the distress of witnessing the misfortune of another person in an event which did not threaten or involve the claimant. Subject to a limited statutory exception in the Fatal Accidents Act 1976, the law did not compensate for the mental anguish and illness flowing from having lost a wife, parent or child or from being compelled to look after a disabled relative. This was not because such consequences could not reasonably be foreseen. It was because of the essential but illusive (Footnote: ¹) concept of proximity or directness.

At page 410H, Lord Oliver said that it was difficult to explain why there was an established exception where injury resulted from “the event of injury to the primary victim” being actually witnessed by the claimant. The answer was to be found in the existence of a combination of circumstances, namely the five elements already mentioned, from which the necessary degree of proximity between the claimant and the defendant could be deduced (page 411D-E).

It is clear from what Lord Oliver had just said at page 410H that, when he used the term “event” in his fifth element, he was referring to “the event of injury to the primary victim” (or the aftermath of it) which was witnessed by the claimant. Thus, the fifth element has to be read as requiring that “there was not only an element of physical proximity to the event [of injury to the primary victim] but a close temporal connection between the event [of injury to the primary victim] and the plaintiff’s perception of it”. That is further confirmed by his reference in the fourth element to the injury needing to arise from “witnessing the death of, extreme danger to, or injury and discomfort suffered by the primary victim”. The word “event” in the fifth element referred back to the event at which the claimant had witnessed the death of, extreme danger to, or injury and discomfort suffered by the primary victim. That is what I have referred to above as the horrific event. The remainder of Lord Oliver’s speech dealt with the relationship between the primary and secondary victims, a question that does not arise in these cases. Lords Jauncey and Lowry concurred without further elaborating on the points of importance to these cases.

In Somerset, the claimant wife claimed damages for psychiatric injury sustained as a result of being told of her husband’s death of a heart attack less than an hour before at the hospital, and identifying his body. The defendant had negligently failed to diagnose the husband’s heart disease for many months before his death. Accordingly, the facts of these cases are similar to the facts of Somerset.

At page 263, Auld J summarised the question he had to decide as being whether the claimant’s involvement within an hour after her husband’s death brought her within the immediate aftermath principle described by Lord Wilberforce in McLoughlin at pages 418-9. Then, at page 267, Auld J said that the claim failed for two reasons advanced by the defendant as follows:

Auld J explained his suggestion that there needed to be an external traumatic event caused by the defendant’s breach of duty, by saying that the immediate aftermath extension was an exception to the general principle established in accident cases that a claimant could only recover damages for psychiatric injury when the accident and the primary injury or death caused by it occurred within his sight or hearing. The two notions implicit in that exception “cautiously introduced and cautiously continued by the House of Lords” were an external traumatic event which immediately caused injury or death, and the claimant’s perception of the event as it happened or shortly afterwards. In that case, there had not been such an external traumatic event “other than the final consequence of Mr. Taylor’s progressively deteriorating heart condition”.

I can say at once that I can find nothing in either McLoughlin or Alcock that imposes the requirement of a traumatic event that is external to the primary victim in the sense of an outside impact or accident, either in the context of the aftermath principle or at all. It seems to me that the judges in McLoughlin and Alcock were doing no more than explaining the principles in the context of cases where such external traumatic events or accidents had in fact occurred. As I have already mentioned, Lord Wilberforce expressly said that cases of nervous shock were not confined to those arising out of accidents. I will return to this question.

In Sion, the Court of Appeal struck out a claimant father’s claim for damages for psychiatric injury caused by an allegedly negligent failure to diagnose his son’s bleeding kidney after a road accident. The father sat at his son’s bedside for 14 days until the son went into a coma and died.

On appeal in Sion, the claimant retracted an acceptance at first instance that the psychiatric injury for which damages were claimed was required to have “resulted from shock – i.e. the sudden and direct appreciation by sight or sound of a horrifying event or events – rather than from stress, strain, grief or sorrow or from gradual or retrospective realisation of events” (page 173 column 2). Staughton LJ seems to have equiparated that submission with one that the claimant did not need to prove an external traumatic event. His decision, on the basis of the medical evidence, was that the claimant did not suffer any shock in the sense of a “sudden appreciation by sight or sound of a horrifying event”, but a continuing process of developing grief.

Peter Gibson LJ agreed with Staughton LJ that the appeal should be dismissed on the ground that there needed to be a “sudden appreciation by sight or sound of a horrifying event, which violently agitates the mind” (Lord Ackner at page 401 in Alcock) rather than “an accumulation of more gradual assaults on the nervous system over a period of time” (page 176). He did not, however, agree with the defendant’s submission that the claim could not succeed because the injuries to or the death of a primary victim in themselves did not qualify as the horrifying event causing the shock needed for a valid claim. He did not agree with Auld J in Somerset where he had “accepted an argument on similar lines”:

Waite LJ agreed with both judgments, without distinguishing between them on an apparent, or at least possible, disagreement between Staughton and Peter Gibson LJJ as to Somerset. Sion is not, therefore, Court of Appeal authority for the requirement suggested by Auld J that there needed to be a traumatic event external to the primary victim. But equally, I do not think it can be taken as Court of Appeal authority approving what I have cited above from Peter Gibson LJ’s judgment (with which Lord Dyson later disagreed in Novo).

Page is not a case that has any real relevance to the issues before this court. The House of Lords coined the phrase “control mechanisms” for the first time and held that they did not apply in a case, such as Page itself, where there was no secondary victim (see, for example, Lord Lloyd at page 197E-F).

In White, police officers who had rescued victims of the Hillsborough disaster failed to recover damages for the resulting psychiatric injuries they had sustained, on the grounds that they did not satisfy the control mechanisms.

Lord Steyn said at page 496G that, despite academic criticism, Alcock was the controlling decision. At page 500C-D, he said that the only sensible general strategy for the courts was to say thus far and no further, treating Alcock as authoritative and leaving any expansion to Parliament. Lord Hoffmann at page 504F thought it was too late to go back on the control mechanisms established in Alcock. Lord Browne-Wilkinson agreed with Lords Steyn and Hoffmann, whilst the other members of the panel dissented.

The Court of Appeal in Walters upheld the judge’s decision to award damages to a mother in respect of the pathological grief reaction that she suffered after waking at her baby’s bedside in hospital when the child was having a fit and then, some 36 hours later, having the child die in her arms after life-support treatment was withdrawn. The defendant had admitted that the child died as a result of its negligent treatment. He ought to have been diagnosed with the acute hepatitis that led to fulminant hepatic failure some days before the fit.

Thomas J had concluded at first instance at [40] (see [17] on appeal) that “[l]ooked at overall and reviewing the factors referred to by Lord Ackner … and Lord Oliver in [Alcock], there was a very strong element of physical proximity to the event which the claimant witnessed, a very close temporal connection between the event and the claimant’s perception of it, the event was horrifying and her appreciation of it sudden in contradistinction to an accumulation of gradual assaults on her mind”. The two main grounds of appeal were that a 36-hour period could not properly be regarded as one horrifying event, and that the claimant had really suffered from a gradual assault of the mind rather than any sudden event.

At [20], Ward LJ said that neither party challenged the judge’s holding that the only real issue had been whether the claimant’s illness had been caused by shock arising from the sudden appreciation by sight or sound of a horrifying event or its immediate aftermath (see also [24]-[25]).

Ward LJ held at [34]-[36] that the entire event was a horrifying one for the claimant mother. By the entire event, he meant the entire 36-hour period starting with the child’s fit. He held that the law permitted a realistic view being taken from case to case of what constituted the necessary event. It is worth noting, however, that, although Ward LJ cited at [29] what Peter Gibson LJ had said in Sion, he did not allude to Auld J’s holding in Somerset that there needed to be a traumatic event external to the primary victim. Nor was any mention made of the gap between at least the start of the negligent failure to treat the primary victim and the onset of the horrifying event.

On the second issue, Ward LJ said that the element of the sudden appreciation of the horrifying event found its place in the definition of shock as an aspect of the proximity which was necessary to establish liability. Lord Oliver had said in Alcock that there had to be both physical and temporal propinquity between claimant and defendant and claimant and the event. Without the sudden and direct visual impression on the claimant’s mind of actually witnessing the event or its immediate aftermath there was no liability. The elements of proximity and causation were closely linked together. In Walters, on the evidence, Ward LJ held that the claimant’s psychiatric condition was caused by shock.

In Novo, the claimant’s mother injured her head and left foot, when a fellow employee tipped a stack of racking boards over her. The defendant employer admitted negligence. After making a good recovery, some three weeks later the mother suddenly collapsed and died in the presence of the claimant daughter. The mother had suffered a deep vein thrombosis and consequent pulmonary emboli, which were caused by the injuries sustained in the accident. The claimant daughter suffered significant post-traumatic stress disorder as a result of witnessing her mother’s death. HHJ Halbert gave judgment for the claimant, but the Court of Appeal (Lord Dyson MR, Moore-Bick and Kitchin LJJ) allowed the defendant’s appeal.

Lord Dyson gave the only judgment. At [3], he said that the defendant’s case was that proximity was lacking because the claimant was not present at the scene of the accident (when the racking boards fell on her mother) and was not involved in its immediate aftermath. The claimant argued that the relevant event was not the original accident, but the collapse and death that resulted from it.

Lord Dyson then reviewed Alcock, White, Somerset, Sion, W v. Essex County Council [2001] 2 AC 592 (Essex), Walters, and Galli-Atkinson v. Seghal [2003] Lloyds Rep Med 285 (Galli-Atkinson) in that order.

Lord Dyson said at [24] that White had explained that the courts should not seek to make any substantial development of the principles; that should be left to Parliament, although the case law showed that some modest development by the courts might be possible.

The issue raised in Novo was whether the mother’s death was a relevant incident for the purposes of the daughter’s claim as a secondary victim [26]. If so, the claim “would not founder on the rock of any of the control mechanisms”. Lord Dyson then pointed out that proximity had been used by Lord Oliver in Alcock in two senses. First, as a legal concept and shorthand for Lord Atkin’s famous neighbour principle. Secondly, proximity was used as one of the control mechanisms to mean the physical proximity in time and space to an event. The correct question in Novo was whether the claimant and defendant were in a relationship of proximity in the legal sense. The difficulty was that, in the context of claims by secondary victims, the concept of proximity depended more on the court’s perception of what was the reasonable area for the imposition of liability than any process of logic. The control mechanisms were the judicial response.

Lord Dyson said this at [29]:

The two reasons that Lord Dyson gave were (i) the daughter would have been able to recover damages for psychiatric illness even if her mother’s death had occurred months, and possibly years, after the accident, and the concept of proximity to a secondary victim cannot reasonably be stretched this far, and (ii) to allow liability would extend the scope of liability to secondary victims considerably further than has been done up to that time. That should only be done by Parliament.

He explained his first reason in an important passage at [30] as follows:

Lord Dyson then said that the judge had been wrong to hold that the mother’s death was the relevant event for the purposes of deciding the proximity question. In the paradigm case of an accident, immediate injury or death is caused to a primary victim and witnessed by the claimant. In such a case, the relevant event was the accident, not a later consequence of the accident.

Lord Dyson explained that Auld J had put the point well in Somerset, when he had said that the immediate aftermath extension had been introduced as an exception to the general principle established in accident cases that a claimant could only recover damages for psychiatric injury where the accident and the primary injury or death caused by it occurred within his sight or hearing. Lord Dyson cited at [11] (and later approved at [33]) Auld J as follows:

Lord Dyson then commented that the claimant daughter would have been able to recover damages as a secondary victim if she had suffered shock and psychiatric illness as a result of seeing her mother’s accident, but that she could not recover damages for the shock and illness that she suffered as a result of seeing her mother’s death three weeks after the accident.

In concluding and dealing with the other authorities, Lord Dyson said that the court was not bound by what Peter Gibson LJ had said in Sion. Peter Gibson LJ had said that he saw no reason why a breach of duty causing an incident involving no violence or suddenness, such as where the wrong medicine was negligently given to a hospital patient, could not lead to a claim for damages for nervous shock, for example where the negligence had fatal results and a visiting close relative, wholly unprepared for what has occurred, found the body and thereby sustains a sudden and unexpected shock to the nervous system. Lord Dyson disagreed. That disagreement is significant for these cases, because the facts that Peter Gibson LJ postulated are close to the circumstances of them. The Court of Appeal seems expressly to have disapproved the concept of a claim for psychiatric injury by a close relative being allowable where clinical negligence caused later fatal results that the relative witnessed.

Essex had, according to Lord Dyson, demonstrated a somewhat more relaxed approach to what constituted the immediate aftermath of an incident. But that was not relevant on the facts of Novo. In Walters the relevant event was one drawn out experience, which was a “seamless tale with an obvious beginning and an equally obvious end ... played out over a period of 36 hours”. In Novo, the primary victim’s injury and death was “certainly not part of a single event or seamless tale”. Lord Dyson explained that: “[t]he question whether the death, being a separate event, was a relevant event for the purposes of a claim by a secondary victim did not arise in Walters”.

Lord Dyson had, as I have said, said that the issue in Novo was whether the mother’s death was a relevant incident for the purposes of the daughter’s claim as a secondary victim. He held it was not because the second consequence of the negligence (the mother’s death) was removed in time from the first consequence (the injury caused by the collapse of the racking boards). That gap in time was fatal to the establishment of legal proximity. What was lacking was, as Auld J had put it, a perception by the claimant of the (first) event as it happened, by presence at or exposure to the scene or to the primary victim so shortly afterwards that the shock of the event as well as of its consequence is brought home to him.

It is not, in my view, entirely clear what Lord Dyson meant by his approval of Auld J’s statement in Somerset that there needed to be an external traumatic event immediately causing the primary victim injury or death. I do not think Lord Dyson can have meant to say that an external traumatic event caused to the primary victim was always necessary, since he did not disapprove Walters where no such external event occurred.

It can be noted, however, that the issue in these cases requires us to identify the relevant event for the purpose of claims by secondary victims of clinical negligence claims, something that was not directly in issue in Novo.

In Ronayne, the claimant husband observed his wife in hospital after undergoing a hysterectomy. Over some 24 hours, the wife deteriorated as a result of the defendant’s negligence. The claimant husband observed two specific incidents. The first was shortly before she underwent emergency exploratory surgery, when she was connected to various machines, and the second was post-operative, when she was unconscious, connected to a ventilator and being administered four types of antibiotic intravenously. The wife was swollen and her husband described her as looking like “Michelin Man”. The main issue in the appeal was whether the events concerned were of a nature to be capable of founding a secondary victim case in that they were in the necessary sense horrifying. The Court of Appeal decided they were not.

Tomlinson LJ approved what Swift J had said in Shorter v. Surrey & Sussex HC NHS Trust [2015] EWHC 614 (QB) at [214]:

Tomlinson LJ’s conclusion is encapsulated in the following at [40]:

Some insight can, in my judgment, be obtained from these passages as to the kind of event that is necessary to found a claim by a secondary victim in a clinical negligence context.

Chamberlain J undertook a meticulous examination of the authorities in deciding Paul’s case on the first appeal. At [63], he noted that the key question was whether Mr Paul’s collapse from a heart attack 14½ months after the allegedly negligent treatment was capable of constituting a relevant event. The first possible reason why not was that the event had to be approximately synchronous with the negligence giving rise to it. Chamberlain J said that there was nothing in McLoughlin and Alcock to suggest that was the case.

Chamberlain J said that Lord Oliver had said in Alcock at page 416 that the “temporal propinquity” required was between the psychiatric injury and “the event caused by the defendant’s breach of duty to the primary victim” (emphasis added), (Footnote: ²) not the breach of duty itself. As Ms Johnson had submitted to him, there was nothing in any of the House of Lords authorities considering the control mechanisms to suggest that a claim for psychiatric injury suffered as a result of witnessing a person’s death or injury caused by (for example) the collapse of negligently erected scaffolding, or electrocution as a result of negligent wiring, would be affected by the date of the negligence. Novo did not suggest that it would. In that case, Lord Dyson had made clear at [29] that the secondary victim would have been able to recover if she had witnessed the accident with the racking boards. There was, decided Chamberlain J, nothing to suggest that the position would have been any different if their collapse had been caused by being negligently stacked months or years beforehand.

Chamberlain J distinguished Novo at [75] as follows:

In Polmear, Master Cook followed the decision of Chamberlain J as to the applicable law. He concluded on the facts of Polmear’s case that the earlier incidents of damage did not rule out a claim by the secondary victim. He said this at [43]:

DJ Lumb reviewed the authorities in Purchase’s case, but his decision was reached before Chamberlain J’s judgment. He concluded at [28] that Novo meant that Mrs Purchase’s claim was doomed to failure. He said that:

DJ Lumb thought that Walters was decided very much on its own facts. The subsequent cases of Novo and Ronayne demonstrated that even a relatively short time period of somewhere between 36 and 54 hours was not in itself sufficient to establish proximity in time. Novo was binding authority that could only be overturned by the Supreme Court or by Parliament changing the law.

These are clinical negligence, not accident, cases. The true question is how the authorities are to be applied to clinical negligence cases where there is a delay between the negligent act or omission and a horrifying event caused to the primary victim by that negligent act or omission.

In essence, there are three situations, which are significant for the decision here. First, there are the accident cases, such as McLoughlin and Alcock, where the negligence and the injury or threatened injury to the primary victim creating the horrific event occur at much the same time. Secondly, cases such as these, where the negligence occurs at an earlier time than the horrific event caused by that negligence. Thirdly, there are the cases, like Novo, where the negligence causes two distinct potentially horrific events separated in time.

Although the five elements were laid down in the first category of case, they are applicable to the other categories of case.

The distinction that has caused so much difficulty and debate is between, on the one hand, the first and third category of cases where the first or only horrific event occurs at much the same time as the negligence and the damage caused to the primary victim, and, on the other hand, the second type of case where there is a gap between these two events. In clinical negligence cases, it is very common for a misdiagnosis to occur at one time and for the death or serious injury to the patient caused by that misdiagnosis to occur much later. But medical negligence is not the only type of case where that can occur. In argument, the court postulated the case of a negligent architect designing a door in a load-bearing wall without specifying an RSJ, causing masonry to fall on a primary victim’s head years later.

Looking at the matter without regard to the authorities, it is hard to see why the gap in time (short or long) between the negligence (whether misdiagnosis or door design) and the horrific event caused by it should affect the defendant’s liability to a close relative witnessing the primary victim’s death or injury that it caused.

Counsel have, as I have explained, suggested different solutions to the problem I have described. Mr Weir for the Paul claimants suggests that the law is that a secondary claimant, who sustains psychiatric injury as a result of witnessing the horrific event suffered by a close relative, can only claim damages for that psychiatric injury if that horrific event is the first manifestation of damage to the primary claimant by the negligence. Mr Bagot for the defendants suggest that the law is that the secondary claimant can only claim damages for psychiatric injury if the horrific event is the damage completing the primary claimant’s cause of action in negligence.

In my judgment, however, these nuanced approaches are distinctions without real differences, even though if they were to be accepted they would, of course, affect liability in particular cases. Each of them would create unprincipled and complex factual disputes as to either when damage caused by the negligence was occasioned to the primary victim or when such damage first manifested itself. There is nothing in any of the cases to suggest that this is the distinction that is to be drawn. I would suggest that it is also illogical to make the liability of a defendant for psychiatric injury caused to a secondary victim depend on whether the primary victim’s cause of action had been complete, or whether that primary victim had sustained manifest damage, before the horrific event undoubtedly caused by the defendant’s negligence. This is even more illogical when one understands that actual injury or damage to the primary victim is not even necessary to found liability to the secondary victim, as Lord Oliver made clear in Alcock.

The approaches suggested by Mr Weir and Mr Bagot are distinctions without differences because the five elements address the legal proximity between the secondary claimant and the defendant, and concern the question of whether that defendant ought, in the eyes of the law, to be liable for the psychiatric injury, even if it was foreseeable. It is, therefore, illogical to drill down into the legal liability and injury caused to the primary victim. What is important is the horrific event itself that caused the secondary victim the psychiatric injury in respect of which the claim is made.

It is important, I think, to start the analysis from Lord Oliver’s five elements. The marital or parental relationship element is satisfied in each of these cases. It is to be assumed in each case here that the injury for which damages are claimed arise from the sudden and unexpected shock to the claimants’ nervous system. It is the third, fourth and fifth elements that are important. They are (iii) the claimant was either personally present at the scene of the accident or was in the more or less immediate vicinity and witnessed the aftermath shortly afterwards, (iv) the injury suffered arose from witnessing the death of, extreme danger to, or injury and discomfort suffered by the primary victim, (v) there was not only an element of physical proximity to the event but a close temporal connection between the event and the claimants’ perception of it.

I have already explained at [33] that the term “event” in the fifth element was referring to “the event of injury to the primary victim” (or the aftermath of it) which was witnessed by the claimant. Thus, the fifth element is to be read as requiring that there is an element of physical proximity to the event of injury to the primary victim, and a close temporal connection between the event of injury to the primary victim and the claimant’s perception of it. Satisfaction of the fourth and fifth elements are not the real dispute here.

The question then is how the third requirement for the claimant to be personally present at the scene of the accident, or more or less in the immediate vicinity, or to witness the aftermath shortly afterwards is to be interpreted in the context of clinical negligence cases. That requirement comes, in part at least, from Lord Wilberforce’s statement in McLoughlin that the fact and consequence of the defendant’s negligence (the accident causing the horrific event) had to be close in time and space to the moment when the claimant was caused the nervous shock.

If one were simply considering these requirements and applying them to the clinical negligence situation, I think one would say that, despite the fact that the horrific event took place later than the defendant’s misdiagnosis: (a) the fact and consequence of the defendant’s negligence (i.e. the event or accident causing the horrific event) was close in time and space to the moment when the secondary victim was caused the psychiatric injury, and (b) the secondary victim was either personally present at the scene of the horrific event or accident or was in the more or less immediate vicinity and witnessed the aftermath shortly afterwards.

The question then is whether the authorities that I have summarised above prevent us reaching that conclusion, which is that broadly advanced by the Polmear and Purchase claimants.

In Walters, the Court of Appeal allowed a claim for damages for psychiatric injury to parents when the negligent diagnosis and failure to treat took place several days before the drawn out horrific event that caused the psychiatric injury. That drawn out event started with the fit that the baby sustained. Nonetheless, the Court of Appeal in Novo held that the question of whether the injuries and the death were separate events had not arisen in Walters, whilst not suggesting that the case was wrongly decided.

The question, therefore, turns on what Novo did decide. It is important first to note that Lord Dyson identified the defendant’s case as being that proximity was lacking because the claimant was not present at the scene when the racking boards fell on the claimant’s mother and was not involved in its immediate aftermath. He decided that the defendant was right and that the gap between the first event (which he seems to have regarded as a potentially horrifying one) and the death three weeks later (which was undoubtedly a horrifying event) was fatal to liability.

Lord Dyson decided at [29] that the defendant’s negligence had two consequences which were separated by three weeks in time. Where there were two (horrific) events caused by negligence, he did not think that the defendant was sufficiently proximate to those suffering psychiatric injury from witnessing the second event. To hold otherwise, he held, would be going too far. Damages for psychiatric illness could not be recovered in respect of consequences witnessed “months, and possibly years” afterwards. That would offend the perception of the ordinary reasonable person, who would not understand why a claimant just missing the aftermath of an accident could not claim, if someone witnessing a second consequence weeks later could claim (see Lord Dyson at [30]). The immediate aftermath extension was as far as the law should go. That was why Auld J had been right in Somerset to say that the claimant needed to perceive the event as it happened by presence at or exposure to the scene or to the primary victim so shortly afterwards that the shock of the event as well as of its consequence is brought home to him. That condition was not satisfied in the clinical negligence case of Somerset itself where the consequence of heart injury and death was many months after the negligence. Lord Dyson thought that the collapse of the racking boards causing injury to the primary victim could have constituted a sufficient horrific event to give rise to a claim by a close relative witnessing it, but what happened three weeks later could not.

The decision in Novo was, as I have said, that the five elements could not be extended to allow a secondary victim to recover damages for psychiatric illness if the horrific event occurred months, and possibly years, after the accident. Lord Dyson did not need to say expressly what would have been the case if the “accident” had occurred some time after the negligence, because he was not dealing with a case in which those were the facts (such as these clinical negligence cases). In my judgment, however, he clearly approved what Auld J had decided in Somerset which was a clinical negligence case. Auld J had made clear that the five elements did not allow recovery for anything that occurred beyond the aftermath of an accident. More specifically he held that there was no qualifying horrific event where the primary victim’s heart condition caused death months after the health authority’s negligence.

In these circumstances, whilst I accept that the actual decision in Walters does not sit easily with Somerset and Novo, I think we are bound by Novo. Novo was decided after full argument about all the relevant preceding cases. I do not think that we can say that it misinterpreted the House of Lords’ authorities. It developed their reasoning even if, as I think, one reading of the five elements as explained in McLoughlin and Alcock would allow recovery by a secondary victim where the negligence and the horrific event caused by it are removed in time.

To answer the question posed by counsel, I do not think it is necessary to decide whether, as Mr Weir submitted, the claimant has to show that the event was the first manifestation of damage caused by the clinical negligence. Neither that formulation nor the one favoured by Mr Bagot (the damage completing the primary victim’s cause of action) is found in the cases. Chamberlain J suggested that the first damage could constitute a relevant event even if it did not occur until 14½ months after the negligence, but that, as I have said is inconsistent with the decision in Novo, by which this court is bound.

I can, therefore, draw the threads together.

First, the five elements required to establish legal proximity in secondary victim cases apply as much to clinical negligence cases as they do to accident cases. The question of what is a relevant horrific event is not dependent either on the completion of the primary victim’s cause of action for negligence or the first manifestation of injury to the primary victim. The primary victim’s cause of action is not the critical thing; there may not always be one. For a secondary victim to be sufficiently proximate to claim for psychiatric injury against the defendant whose clinical negligence caused the primary victim injury, the horrific event cannot be a separate event removed in time from the negligence. If the negligence and the horrific event are part of a continuum as seems to me the best possible explanation of Walters, there is sufficient proximity. It may be that the negligence was continuing in Walters at the time the 36-hour shocking event began. Either way, Novo is binding authority for the proposition that no claim can be brought in respect of psychiatric injury caused by a separate horrific event removed in time from the original negligence, accident or a first horrific event.

I am, for the reasons I have given, unable to agree with Chamberlain J’s conclusion at [75]. In my judgment, Novo does preclude liability in the circumstances of these cases, even where a horrific event is the first occasion on which any damage is caused to the primary victim.

I would allow the appeals in Paul and Polmear, and dismiss the appeal in Purchase.

I have, as I have already said, reservations about whether Novo correctly interprets the limitations on liability to secondary victims contained in the five elements emerging from the House of Lords authorities. Subject to hearing further argument, therefore, I would be prepared to grant permission to the claimants to appeal to the Supreme Court, if sought, so that it can consider the important issues that arise in this case.

I agree with the conclusion and reasoning of the Master of the Rolls, but because of the importance of the issue I will state in summary form my reasons for doing so.

The essential feature of these three cases (as pleaded) is that the “shocking event” – that is, the death of the primary victim which (or the immediate aftermath of which) was witnessed by the claimant and caused the psychiatric illness complained of – occurred appreciably after the omissions which constitute the defendants’ negligence. The issue is whether that means that the injury to the claimant is not sufficiently “proximate” to the breach of duty.

That question is not directly answered by the decisions of the House of Lords in McLoughlin, Alcock and White. They were concerned with cases where the death of the primary victim, which (or the immediate aftermath of which) was the shocking event witnessed by the secondary victim, was broadly contemporaneous with the breach of duty; and it was unnecessary to consider what the position would be if it had occurred some time later. However, I find it hard to see a principled reason why there should be the requisite proximity in the one kind of case but not the other. The arbitrariness of the distinction is illustrated by the example given by Peter Gibson LJ in Sion: why should the doctor who negligently prescribes a fatal medicine be liable to the secondary victim if the patient takes it, and dies (in the requisite shocking circumstances) straightaway, but not if they do so only a few days or weeks later? As the Master of the Rolls demonstrates, Lord Oliver’s references in Alcock to the need for “physical and temporal propinquity” are not directed to the relationship between the breach of duty and the shocking event but rather to the need for the claimant to be close in space and time to the shocking event.

It follows that if the point were free from authority I would be minded to hold that on the pleaded facts the Claimants in all three cases should be entitled to recover. I do not think that recognising the necessary proximity in such cases would be contrary to the “thus far and no further” approach taken in White. It would not involve going beyond the elements established in Alcock: rather, it would represent their application in a different factual situation.

The question thus is whether we are prevented from reaching that conclusion by any decision of this Court. For the reasons given by the Master of the Rolls, the decisions in Sion, Walters and Ronayne are not authoritative on the present issue. The difficulty, however, is Taylor v A. Novo. I have not found it easy to identify the precise ratio of that decision. There are some parts of the reasoning in paras. 29 and 30 of Lord Dyson’s judgment that suggest that he saw the crucial feature of the case as being that there had been an initial “accident” which could in principle have constituted a shocking event, if witnessed, and that that meant that any subsequent such event could not be regarded as sufficiently proximate: that was Chamberlain J’s analysis. But I have come to the conclusion that if the judgment is read as a whole the ratio is wider than that. At para. 33 Lord Dyson said that “it follows from what I have said above” that the reasoning of Auld J in Taylor v Somerset Health Authority was correct, and he rejected Peter Gibson LJ’s contrary observations in Sion on the basis that they were obiter dicta. The essence of the reasoning which he was thereby endorsing was that there could be no liability for psychiatric injury to the claimant caused by witnessing her husband’s death “long after the negligence which had caused it”, as opposed to injury caused by a traumatic event occurring at the time of the breach of duty (see the passage quoted at para. 11). Against that background, the fair reading of paras. 29 and 30 seems to be that the ultimately decisive feature was simply that there had been an interval of time between the breach of duty, whether or not it occasioned any injury at the time, and the shocking event. In Taylor v A. Novo itself the interval was three weeks, but the principle must be the same whatever the interval, provided it is not part of the same sequence of events as in cases of the Walters kind.

I should add that I entirely agree with the Master of the Rolls’ rejection at paras. 81-83 above of the approaches contended for by Mr Weir and Mr Bagot based on when an injury first manifested itself or a cause of action accrued, which seem to me both unprincipled and unworkable.

On that basis none of the present claims can succeed. I would accordingly agree with the Master of the Rolls’ proposed disposal of these appeals. My strong provisional view, like his, is that the issues raised by them merit consideration by the Supreme Court.

I have read the judgments of the Master of the Rolls and the Vice President of the Civil Division of the Court of Appeal, Lord Justice Underhill, and agree with each of them.