Chapter 5: Medical evidence: the testimony of the dead

¶What was already known 2.5.1 Evidence relating to the cause of death was central to the 95 'mini-inquests' conducted by the South Yorkshire West District Coroner, Dr Stefan Popper, alongside the summaries from West Midlands Police about the place of death. Three features recurred across the medical and pathological evidence given to the inquests. 2.5.2 First, traumatic asphyxia was a central feature, in most cases as the underlying cause of death. Second, in their evidence the pathologists presented a common account that consciousness would have been lost within a matter of seconds and irreversible brain damage would have occurred in minutes. 2.5.3 Taken together, these features presented an unvarying picture of a uniform, rapid process that led inevitably to death once an irresistible pressure had built up within the central pens. As such, as discussed in Chapter 10, it underpinned the Coroner's decision to impose a 3.15pm cut-off on evidence presented at the generic stage of the inquests. 2.5.4 The third recurring feature was the emphasis attached to alcohol, as the blood alcohol level of the deceased was read to the court at the start of each 'mini-inquest' and immediately reported in the media. 2.5.5 The disclosed documents add significant new information on each of these crucial aspects of the medical evidence. The first part of this chapter considers the evidence available from systematic review of the pathology reports. The second part highlights the significance of the Coroner's exceptional decision to take blood alcohol samples from the deceased, and how the results were presented.

¶Pathology 2.5.6 The investigation into the deaths included a post mortem examination of each body. This was carried out under the jurisdiction of the Coroner for the district in which the deaths occurred, in this case the South Yorkshire West District Coroner, Dr Popper. It is usual to conduct post mortem examinations when deaths occur that cannot be attributed reliably to natural causes. 2.5.7 Yet Dr Popper's contemporaneous notes indicate that this was not a foregone conclusion: 'I considered the need for post mortem in these cases, bearing in mind that visual inspection indicated that a probable conclusion would be Traumatic Asphyxia and bearing in mind that many of the deceased were young people'.1 1. File note by Dr Popper, 16 April 1989, SYC000001360001, p245.

Chapter 5 Medical evidence: the testimony of the dead — 159

160

¶2.5.8 It is instructive that as early as the morning of Sunday 16 April 1989, within 24 hours of the disaster and before any post mortems had been conducted, the Coroner had surmised that the probable cause of death was traumatic asphyxia for all 94 people who, at that time, had died. It was a conclusion of sufficient certainty that he questioned the need for post mortem examinations.

¶Arrangements for post mortems 2.5.9 Yet, on balance, Dr Popper decided that post mortems were required and all would take place at the Medico-Legal Centre in Sheffield, in as short a time as practicable. To expedite the process, additional pathologists attended the Medico-Legal Centre, and nine pathologists carried out 94 post mortem examinations over two days.

¶2.5.10 Two people died later, one following two days in hospital and a second in 1993, after being in a persistent vegetative state since the disaster. These two post mortems were carried out by different pathologists, the latter under the jurisdiction of the West Yorkshire Coroner, as he had died in Airedale Hospital.

¶2.5.11 The arrangements for the post mortem examinations were in accordance with legal requirements and with standard practice, although to a demanding timescale. The reason for this haste is not clear from the documents. Nor is the reason for the other outstanding feature, the unusual direction that blood be taken from all of the deceased at post mortem to determine a blood alcohol level.

¶2.5.12 From subsequent statements2 it is clear that this directive was decided by Dr Popper before the post mortem examinations began, and it is clear from the post mortem records that the directive was followed in each of the 94 post mortems on those who died, regardless of age. In addition the documents confirm that a blood alcohol level was estimated in the 95th, a boy of 14 who died in hospital two days after the disaster, using a sample taken previously.3

¶2.5.13 Blood alcohol levels are routinely checked in those driving or piloting motor vehicles, railway trains, ships and aircraft involved in fatal incidents, but not in mass disaster victims.

¶Post mortem reports 2.5.14 The Panel regards the records of the post mortem examinations as confidential to the family concerned and not for public disclosure. Its terms of reference, however, require a report on the overall content of material shared with the Panel. All post mortem reports were scrutinised in detail by a medically qualified Panel member, and are described in aggregate here.4 The results show some striking features, considered under four headings: cause of death; traumatic asphyxia and venous compression; cerebral oedema; and implications of post mortem reports.

Inquest transcript, opening statement by the Coroner, 18 April 1990, SYC000109270001, p31. 3. Inquest transcript, 1 May 1990, SYC000109960001, p8. 4. In view of the specialist nature of some of the pathology the overall findings were discussed with an independent expert forensic pathologist, and the Panel is grateful for his helpful advice.

161

¶Cause of death 2.5.15 As required in England and Wales, recording of cause of death allows for a chain of up to three conditions, the first of which is the 'immediate cause of death' and the last being the 'underlying cause of death'. They might be common if only one cause is listed. It is also possible to note 'associated conditions' which contributed, but did not lead directly, to death.

¶2.5.16 The immediate cause of death was given as traumatic asphyxia5 in 68 cases and as crush asphyxia in 14. Most forensic pathologists would regard the terms 'traumatic asphyxia' and 'crush asphyxia' as interchangeable, although some may seek to draw a distinction between a single impact or compression causing traumatic asphyxia, and a more gradual compression causing crush asphyxia.

¶2.5.17 This approach was taken, for example, by Mr James Wardrope, Accident and Emergency Consultant at the Northern General Hospital Sheffield, and his colleagues in describing the outcome of treatment of those admitted to hospital following the disaster.6

¶2.5.18 However, it is clear from the answers given repeatedly to questions during the inquests that the pathologists drew no such distinction and regarded the two terms as synonymous. For example, Professor Alan Usher, the senior pathologist at the Medico-Legal Centre, was explicit in his evidence: 'Traumatic asphyxia, which we talked about yesterday, is sometimes crush asphyxia for obvious reasons and some of the pathologists have used that term and some have used traumatic asphyxia. There is no difference'.7

¶2.5.19 Other immediate causes of death recorded were inhalation of stomach contents8 (6), inhalation of stomach contents together with traumatic asphyxia (1), respiratory failure9 (2), cerebral anoxia10 (1), pyelitis11 and bronchopneumonia12 (1), cardiorespiratory arrest13 (2), and shock and haemorrhage14 (1).

¶Traumatic asphyxia 2.5.20 The underlying cause of death shows an even greater preponderance of traumatic asphyxia (73) and crush asphyxia (17) – 90 in total (one jointly with inhalation of stomach contents). Of the remaining six, the underlying cause of death was given as inhalation of stomach contents in three, traumatic pulmonary contusions15 in two, and transection of the aorta16 in one. In four of these six where neither traumatic nor crush asphyxia was the underlying cause, one or other was given as an associated cause contributing to death.

Asphyxia is a lack of oxygen in the body, often due to a problem with breathing. Traumatic asphyxia is a lack of oxygen due to compression of the chest preventing breathing, and often obstructing the blood flow back to the heart. 6. Wardrope J, Ryan F, Clark G et al. The Hillsborough Tragedy. British Medical Journal 1991; 303: 1381-1385. Available on the Panel's website at HOM000038420001. 7. Inquest transcript in respect of Stephen Francis O'Neill, 14 May 1990, SYC000109290001, p7. 8. Obstruction of breathing due to the effect of stomach contents on the airways if regurgitated and inhaled. 9. Inability of the lungs to function adequately, particularly to supply oxygen to the bloodstream. 10. Lack of oxygen affecting the brain. 11. Inflammation of the upper part of the urinary system, particularly due to infection. 12. Infection of the lungs and the airways leading to them. 13. Cessation of heartbeat and breathing, for example due to the brain ceasing to function. 14. Blood circulation inadequate due to loss of blood. 15. Bleeding into the substance of the lung due to injury, for example from pressure on broken ribs. 16. Complete division of the main blood vessel leading from the heart.

162

¶2.5.21 Thus in only two cases does neither traumatic nor crush asphyxia appear on the certificate. In these, respiratory failure due to traumatic pulmonary contusions associated with fractured rib and pulmonary lacerations was recorded. Even when the cause of death was certified as shock and haemorrhage due to transection of the aorta, one of the most rapid causes of sudden death, traumatic asphyxia was given as an associated cause contributing to death.

¶The 'mini-inquests' 2.5.22 The issue of traumatic asphyxia recurred consistently during the preliminary hearings into each individual death ('mini-inquests'). The pathologist who carried out the post mortem was invited to agree that, as a result of traumatic asphyxia, loss of consciousness would have occurred rapidly, within seconds, and that death would have followed within a few minutes at most.17

¶2.5.23 In each case, the pathologist accepted this interpretation. This was emphasised to the families as a matter of comfort, but it also established an unvarying pattern of death, a matter of importance to the Coroner in his approach to the inquests. Subsequently, when aspects of the conduct of the inquests were challenged through Judicial Review, he prepared a statement of evidence explaining his decisions.

¶2.5.24 His initial draft stated: 'In every one of the 95 cases the uncontested evidence of the pathologists was that the pathological cause of death was traumatic asphyxia and that within a matter of seconds the individual would have been unconscious and unaware of anything further and would have died within a matter of minutes thereafter'.18 Subsequently, 'In every one of the 95 cases' was amended to 'in the majority of the cases' and presented as the final version.19

¶2.5.25 While this insistence on a single unvarying pattern of rapid death may have been motivated, at the time of the mini-inquests, by a desire to ease the emotional burden on relatives, it was a crucial factor in the imposition of the 3.15pm cut-off. Consequently, as shown in Chapter 10, evidence concerning events after that time was not considered at the generic stage of the inquests, a cause of significant distress to relatives.

¶2.5.26 The Coroner argued that the outcome for each of those who died was determined entirely by events before 3.15pm, and that no new significant event could have intervened in the chain of causation of death beyond that time:

¶As a marker I picked the arrival of the first ambulance on the pitch which was timed at

¶3.15pm because on the overwhelming pathological evidence available to me, by that time permanent irreversible damage would have already occurred.20

¶2.5.27 This view of the rapidly fatal and irreversible nature of traumatic asphyxia also influenced LJ Taylor, who concluded that the potential impact of the emergency services was limited: 'in view of the nature and extent of the crushing, the time when police rescue began and the pathetically short period for which those unable to breathe could survive, it is improbable that quicker recourse to the emergency services would have saved more lives'.21

For example, at the inquest in respect of Peter McDonnell, 20 April 1990, SYC000109440001, p7. 18. Draft Affidavit by Dr Popper, undated, SPP000002120001, p7. 19. Affidavit by Dr Popper, undated, SYC000001290001, p12. 20. Affidavit by Dr Popper, undated, SYC000001290001, p14. 21. Interim Report of the Inquiry into the Hillsborough Disaster (Taylor Report), HOM000011140001, p59.

163

¶2.5.28 However, the portrayal of an unvaryingly rapid and inevitable death was not supported by the post mortem findings in a substantial number of cases. Two principal findings emerged from the Panel's review of all of the post mortem reports.

¶Traumatic asphyxia and venous compression 2.5.29 In an expert medical opinion provided for the Judicial Review of the Inquests, Dr Iain West, a consultant forensic pathologist, was critical of key aspects of the eight post mortem reports on which he had been invited to comment.22

¶2.5.30 In particular, he stated that a distinction should be drawn between 'classic' traumatic asphyxia, where a sudden rise in venous pressure results in rapid cessation of circulation and a high probability of death, and asphyxia due to suppression of breathing through mechanical compression of the chest wall without venous obstruction.

¶2.5.31 This takes longer to develop and is associated with a greater likelihood of rescue from mechanical compression of someone partially asphyxiated but still alive. These distinct conditions present different appearances at post mortem. The venous compression characteristic of 'classic' traumatic asphyxia results in intense congestion and a deep purplish-blue skin colouration (cyanosis) with many small (petechial) haemorrhages, occurring over the head, neck and upper chest.

¶2.5.32 Asphyxia without venous compression may result in cyanosis and a few fine petechial haemorrhages, particularly over the head, neck and extremities, but not the marked pattern restricted to the upper part of the body and associated with congestion that is caused by venous compression.

¶2.5.33 Dr West found no evidence of 'classic' traumatic asphyxia in three or four of the eight reports that he scrutinised. He concluded that:

¶it is impossible to state purely from the medical point of view that a number of the young men that I have indicated above could not have been alive at 3.15pm. Those dying as the result of anoxic damage consequent to their chests being crushed could well have survived for a much longer period only to die subsequently from the effects of irreversible anoxia.

¶Access to post mortem records 2.5.34 Dr West had access to only eight post mortem records. With access to all post mortem records, the Panel was able to review the entire set against these criteria. In 15 of the post mortem records there is a clear description of the findings of 'classic' traumatic asphyxia with venous obstruction, and in a further 25 the description suggests probable venous obstruction.

¶2.5.35 In 28, however, the findings described clearly do not support the occurrence of 'classic' traumatic asphyxia with venous obstruction, and in a further 16 a significant degree of venous obstruction is unlikely from the description given. (In 11 the appearances were insufficiently clearly described to decide, while in the 96th, death occurred after a prolonged period in hospital by which time the initial changes had reversed.)

Written opinion of Dr I West, Department of Forensic Medicine, Guy's Hospital, 20 August 1992, SYC000001280001, pp66-71.

164

¶2.5.36 The occurrence of a substantial proportion with evidence of this different form of asphyxia calls into question the medical evidence presented to the inquest of a single unvarying pattern of death due to traumatic asphyxia. 'Classic' traumatic asphyxia results in cessation of the blood circulation unless relieved, because the venous obstruction prevents blood returning to the heart. In contrast, those without significant venous obstruction and circulatory arrest are likely to have survived for a significantly longer period.

¶2.5.37 Had their chest compression been relieved during this period, for example by removal from the pens, resuscitation of a partially asphyxiated individual with a continuing heartbeat would have been a very different proposition from resuscitation of someone who had already suffered cardiac arrest, and significantly more likely to lead to a successful outcome.

¶2.5.38 Importantly, a person in this condition would also have been vulnerable to further potentially fatal asphyxia from a new cause, such as airway obstruction from being positioned on their back or from inhalation of stomach contents.

¶Cerebral oedema 2.5.39 The Panel's scrutiny of all the post mortem reports showed a second feature that casts significant doubt on the notion of a single, unvarying mode of death: the description in some of cerebral oedema.

¶2.5.40 Cerebral oedema is a swelling of the substance of the brain due to fluid that has left the bloodstream and accumulated in and around the cells of the brain. In this context it occurs as an effect of shortage of oxygen in the blood. The accumulated fluid compresses the substance of the brain, gradually affecting brain function, and increases its weight.

¶2.5.41 Because the brain is almost totally enclosed in the rigid bony cavity of the skull, if sufficient swelling occurs it results in parts of the brain being forced by the increased pressure through the main opening at the base of the skull where the spinal cord passes. This is described as 'coning'.

¶2.5.42 The appearance of cerebral oedema was clearly described in 31 of the post mortem records, and was sufficient to cause coning in 16 of these. In a further ten, coning was described but the brain was not recorded as enlarged. As the significance of this description is not clear, these have been disregarded, as have two in whom coning was associated with, and probably due to, bleeding around the brain.

¶Significance of cerebral oedema 2.5.43 The importance of this finding is that cerebral oedema takes significant time to develop, and longer to progress to the point at which coning occurs. During this time, the blood circulation to the brain must have continued, since once it ceases, cerebral oedema cannot develop further as no more fluid is being supplied to the brain.

¶2.5.44 Cerebral oedema is not described in cases of immediate complete asphyxiation, for example full strangulation, hanging or rapidly fatal traumatic asphyxia with venous compression, because the circulation stops within a few minutes, before detectable cerebral oedema can accumulate. It is found, consistently, however, in people who have survived for prolonged periods after partial strangulation or hanging, and among survivors of traumatic asphyxia, when there has been sufficiently severe asphyxia to cause unconsciousness through lack of oxygen but not sufficient to cause immediate circulatory arrest.

165

¶2.5.45 It is notable that not only was cerebral oedema described at post mortem in the person who survived for two days, it was also found in the most seriously ill individuals who were admitted to intensive care after the disaster and subsequently survived. These patients were cared for by Dr Frank Ryan, Consultant Physician at the Northern General Hospital Sheffield, who observed that:

¶[w]ithin ten or fifteen minutes, a total of 13/14 patients appeared to exhibit a very similar syndrome. They were either unconscious already or were partially conscious, appearing confused and bewildered ... All of them went on to develop status epilepticus.23 This, under the circumstances, was life-threatening ... It was my opinion that every patient who was fitting had cerebral oedema and they should all be ventilated and receive intensive care.24 2.5.46 The finding of cerebral oedema at post mortem was raised during some mini-inquests. Generally it was attributed by the pathologist giving evidence to the effects of particularly severe asphyxia, without comment on the time necessary for its development. 2.5.47 The fullest account occurred in evidence given by Professor Alan Usher.25 He observed that: 'The signs of traumatic asphyxia both internally and externally were quite marked'. In fact there were only a few petechial haemorrhages described and no upper body cyanosis or venous congestion. He continued:

¶In this case there was swelling of the brain and coneing [sic] of the hind brain and I thought this was sufficiently significant to include it in the cause of death ... when you insult the brain, in almost any way, by shaking it about in the head or by not supplying it with oxygen, it has one reaction and that is to swell and, in this case, it swelled inside the closed box of the skull and protruded down through an opening which the spinal cord goes down into the spine through and that caused pressure on the vital areas of the brain which would kill very rapidly ... I think that once the pressure was onto his chest, then he would have become unconscious ... in a very short time indeed – a matter of seconds – probably between 10 and 20 seconds, so whatever happened subsequent to that he would not feel. 2.5.48 He was asked: '... once that unconsciousness sets in, the swelling of the brain, as I understand it, is very rapid. It is not a slow process, it is a very rapid process?' His answer was 'Yes'. He was then asked: 'So that there would have been nothing that could have been done by the time this young man had, for example, been taken onto the pitch minutes after the crushing?' He responded: 'No, as I say, had he survived he would almost certainly have been physically disabled because of brain damage'. 2.5.49 These responses omit the most important aspect – the length of time that cerebral oedema takes to develop. It is correct to say that cerebral oedema may begin to develop soon after the onset of the shortage of oxygen affecting the brain, and it is also correct to say that after it has progressed to the point of coning of the lower part of the brain severe damage will occur that may be rapidly fatal. 2.5.50 The crucial point, however, is that progression of cerebral oedema from its first onset to the point of coning is not a rapid process. It takes significant time to develop. It is regrettable that this was not brought to the Coroner's attention in response to this questioning. Nor was the occurrence of cerebral oedema in such a large number of those who died, many with coning, which was not recorded as part of the cause of death.

Status epilepticus is the occurrence of an uninterrupted series of convulsions or 'fits'. 24. Dr FP Ryan, 'Summary of my experience with the Hillsborough Tragedy', 19 April 1989, FPR000000110001, p2. 25. Inquest transcript in respect of Carl David Lewis, 3 May 1990, SYC000110140001, pp6-7.

166

¶2.5.51 These individuals must have survived for a period sufficient for cerebral oedema to develop to the onset of coning. During this period the circulatory system would have functioned, with at least some continued respiration to maintain the circulation, or cerebral oedema would have stopped developing. 2.5.52 Although these individuals were unconscious and in imminent danger of death from asphyxiation, it is difficult to conclude with certainty that rescue and resuscitation attempts during this period were irrelevant to their survival. As shown in Chapter 4, people did survive cerebral oedema due to partial asphyxiation, the majority without lasting neurological damage. It is also difficult to maintain that no new event could have occurred to precipitate death in somebody in this condition who might otherwise have survived.

¶Implications of post mortem reports 2.5.53 Taken together, these features of the post mortem reports not only confirm Dr West's conclusion that some individuals died later than 3.15pm, but more significantly they imply that there remains considerable doubt concerning the assumption that once maximum compression had occurred in the central pens the outcome was predetermined for all who died, and that no new factor could have intervened in the chain of causation of death. 2.5.54 This was a key part of the Coroner's reasoning leading to his determination of the 3.15pm cut-off, and his decision not to systematically consider evidence post 3.15 at the generic inquest. In reviewing this decision Lord Justice Stuart-Smith defended the Coroner's approach:

¶It should be noted that the Coroner did not say that all those who died did so before

¶3.15, or that the medical evidence was to this effect … Nor did he say that all those who became unconscious subsequently died. The evidence was that it was only those people whose chests were in a state of permanent fixation as a result of the crush for four to six minutes, so that they could not breathe at all for that time, whose condition was irreversible.26 2.5.55 The evidence that in some people respiration and circulation continued for a significant period, from the release of compression until they died, clearly challenges the assumption that their condition was irreversible. Some in this condition survived after treatment in an intensive care unit, while some died at the ground. It is likely that what happened to them during that period played a major part in determining the outcome. 2.5.56 This point was clearly illustrated by Dr James Burns, a forensic pathologist who reviewed a single post mortem report at the request of a bereaved family:

¶... in the case of a person removed from the enclosure at, say, 3.10pm and who was unconscious, but not brain dead, and was then placed in a position other than the correct 'recovery' position, or who, having been placed in the correct position, such a position was not maintained, a feared consequence, an inadequate airway, may well have produced a prolonged state of unconsciousness, with death eventually occurring at a much later time than 3.15pm.27 2.5.57 Dr Burns and Dr West had access only to a few post mortem reports of those who died at Hillsborough. A systematic review of all disclosed reports shows that 28 clearly had no signs of venous compression and that traumatic asphyxia, at least in its 'classic' form, was not an appropriate description of the cause of death. 26. Scrutiny of evidence relating to the Hillsborough football stadium disaster, by Lord Justice Stuart-Smith,

¶HOM000045010001, p48. 27. Letter from Dr James Burns to Brian Thompson & Partners solicitors, 11 March 1991, SYP000096240001, pp46-47.

167

¶2.5.58 There is also evidence in 31 cases that circulation and respiration continued for a period of time, sufficient in 16 for cerebral oedema to progress to its fullest extent. This renders untenable the notion, first voiced by the Coroner before any post mortems had been carried out, of a single, unvarying cause and pattern of death.

¶2.5.59 Further, it challenges the Coroner's conclusion that nothing that happened after release from the pens could affect survival. On the one hand, basic first aid aimed at clearing and maintaining an airway might have kept people alive long enough for them to be treated in hospital, as in the case of those people admitted to hospital who subsequently developed cerebral oedema, all but one of whom survived.

¶2.5.60 On the other hand, placing an unconscious person flat on their back, as is known to have happened in some cases, potentially would lead to further asphyxia from obstruction of the airway. Unless promptly relieved this would most likely prove fatal, without leaving any further post mortem signs in addition to those already expected from asphyxia due to restriction of breathing by chest compression.

¶Blood alcohol measurement Introduction 2.5.61 The emphasis placed at the opening of each mini-inquest on reading out the blood alcohol level of the deceased person, estimated from a blood sample taken from each of the deceased under the direction of the Coroner, was a recurring feature of the inquests. Except for the first two mini-inquests, when it appears to have been overlooked, this was the first evidence presented to the jury at each mini-inquest. Inevitably, the impact was to suggest that alcohol was central to the disaster and relevant to each death under consideration.

¶2.5.62 Media coverage reinforced this impression, as did repeated comments by police sources, but the scene was set at the mini-inquests. Two witnesses who gave evidence at the generic stage of the inquests returned to this theme. Dr Alexander Forrest, a forensic toxicologist at Sheffield's Royal Hallamshire Hospital who had carried out the analysis of the blood samples removed at post mortem, suggested that even modest blood alcohol levels might be associated with an impaired reaction to novel situations.

¶2.5.63 Dr Jonathan Nicholl, an epidemiologist commissioned by the Coroner to investigate a possible relationship between blood alcohol levels and the time of entry of the deceased into the ground, claimed that those who entered later were more likely to have a raised blood alcohol level.28 In his final summing up, the Coroner reinforced the impression that alcohol was a relevant factor, drawing on the evidence of Dr Forrest and Dr Nicholl and linking them with impressionistic, subjective accounts by police officers of unspecified, intoxicated fans.

¶Blood alcohol levels 2.5.64 Although the results of blood alcohol estimations were read out at the Coroner's direction during the initial stage of each individual mini-inquest, after the first two, and reported daily in the press coverage of the inquests, there was no attempt to assess whether the results had any significance for the individual or for the occurrence of the disaster.

¶undated, SYC000000960001, pp21-32.

168

¶2.5.65 Media coverage related the results to the 'drink-drive limit' of 80mg of alcohol per 100ml of blood. Although only 15 of those who died had a blood alcohol reading above this threshold, each was reported as if it was a significant factor in the context of the disaster. Yet there was no suggestion that any of the deceased over this limit had driven, or intended to drive, a vehicle.

¶2.5.66 Nor was there any systematic consideration of what relevance there might be for those attending a social occasion, a football match, of a drink-drive limit that is set to prevent people driving who are not visibly or behaviourally intoxicated but whose delayed reactions and coordination would impair control of a motor vehicle at speed.

¶2.5.67 The blood alcohol estimations were carried out by Dr Forrest. In evidence to the preliminary proceedings of the mini-inquests, he commented on the significance of blood alcohol levels for individuals:

¶People do vary enormously in their response to alcohol. Objective tests by the Road

¶Traffic Research Laboratory and also studies on the rate of accidents after people have particular amounts of alcohol in their blood, show that people with a blood alcohol concentration of between 20mg to 40mg of alcohol/100ml of blood are perhaps somewhat impaired in their ability to respond to a novel situation.

¶I have seen individuals with blood alcohol concentrations of 200mg to 300mg of alcohol/100ml of blood who on cursory examination would appear to be perfectly sober and to be able to conduct a normal conversation ...

¶On the other hand, I have seen a young man from this part of the world with nothing else to show for it who was dead with a blood alcohol concentration of less than

¶80mg of alcohol/100ml of blood.29

¶2.5.68 Neither Dr Forrest nor the Coroner, in reiterating the remark about the Road Traffic Research Laboratory tests, made the obvious point that they measured response times to very rapidly changing situations, relevant to drivers in charge of a motor vehicle, but not to pedestrians attending a leisure event.

¶The Jones Report 2.5.69 The measurement of blood alcohol and its significance were the subject of a report prepared by Professor Wayne Jones, an international authority on alcohol testing, commissioned for the private prosecution.30 The report criticised several technical aspects of the testing, including the sites from which blood was taken, the failure to obtain confirmatory samples from the bladder or eye, and the analytical technique.

¶2.5.70 Professor Jones disagreed with Dr Forrest that a blood alcohol level of 20mg per 100mg was of any significance. Dr Forrest had suggested in evidence that this level might be found in someone who had drunk a pint or two of beer the previous night, but the Jones Report pointed out that the rate of metabolism would have cleared alcohol consumed the previous night. Professor Jones also emphasised that any post mortem level of less than 50mg per 100ml is of dubious significance and likely to be an artefact due to post mortem changes.31

Evidence of Dr Forrest at the Hillsborough Inquest, 18 April 1990, SYC000109270001, p71. 30. 'Review and Opinion, Preliminary Report', by Professor AW Jones, 18 April 1990, FAM000000010001, pp3-11. 31. Fermentation due to bacteria can produce alcohol in the body after death.

169

¶Irrelevance of drink-drive limit 2.5.71 The Jones Report was also critical of the use of the drink-drive limit as if it had relevance to attendance at a football match. The report is quite clear: 'the insinuation that many of the victims were drunk (BAC>80mg/dL) at the time of the disaster and thus too impaired through drink to respond to a novel situation and that this played some role in their death is unjustified'.

¶2.5.72 If all results of the blood alcohol testing are reclassified according to levels indicated by Professor Jones, a clear picture emerges. Of the 95 individuals who died as an immediate result of the disaster, a total of 68 had undetectable (55), or insignificant and probably artefactual (13) levels of alcohol.

¶2.5.73 A further 12 had levels compatible with minor social disinhibition, and nine had some impairment of rapid responses, therefore unable to drive legally. Only six of the 95 had levels at which they may have been expected to show signs of being intoxicated.

¶2.5.74 In marked contrast to the prevailing assumption originating at the inquests, and widely promulgated through public statements made by senior South Yorkshire Police officers and published in the press, this notably modest pattern of alcohol consumption would bear comparison with any social, sporting or leisure occasion, and clearly endorses LJ Taylor's conclusion that drunkenness played no part in the disaster.

¶The Nicholl Report 2.5.75 The restrained nature of this overall pattern of alcohol consumption among spectators at a football match was not considered or explored at the inquests. In fact, Dr Jonathan Nicholl of Sheffield University was commissioned by the Coroner to write a report investigating a possible association between time of entry to the ground and blood alcohol level among those who died, suggesting that latecomers with higher blood alcohol levels may have been significant in what developed.32

¶2.5.76 Dr Nicholl presented a summary of his report at the inquests which, he claimed, confirmed an association between later entry to the ground and raised alcohol levels.33 Dr Nicholl did not make a link between the levels involved and the occurrence of the disaster. Yet it is clear from the Coroner's summing up that he placed emphasis both on the 'fall off in quality and manoeuvrability' in those over the drink-drive limit, and on Dr Nicholl's finding that those who had entered the ground after 2.30pm were more likely to have a raised blood alcohol level.

¶2.5.77 Disclosure of the original data analysed by Dr Nicholl, however, casts substantial doubt on his findings.34 In order to demonstrate this, the Panel has both replicated Dr Nicholl's original analysis and also reanalysed the original figures, revealing six significant problems with his report.

¶2.5.78 The first is Dr Nicholl's treatment of the data on time of entry. He established five categories, as well as an 'unknown entry' group who were excluded from analysis. There were three groups known to have entered via a turnstile, either before 2.30pm, between 2.30pm and 2.47pm, or after 2.47pm. There was a group known to have entered via Gate C, after 2.47pm when the gate was first opened. Another group were those whose route

'Hillsborough – Association between time of entry to the ground, age and alcohol consumption', by Dr JP Nicholl, undated, SYC000000960001, pp21-32. 33. Evidence of Dr Nicholl at the Hillsborough Inquest, 14 March 1990, SYC000109160001, pp4-33. 34. Data used by Dr Nicholl, SYC000000960001, pp5-20.

170

¶of entry via a turnstile or Gate C was unknown, but their entry time was believed to be 'probably' after 2.30pm. These groups are shown diagrammatically in Figure 5. The group whose entry route was unknown and who 'probably' entered after 2.30pm clearly constitute an awkward category for the analysis.

¶Figure 5: Route and time of entry of those who died Each oval represents one person. Note route of entry of middle group unknown, time 'probably after 2.30pm'. (Excludes 13 whose route and time are unknown.)

¶2.5.79 Dr Nicholl's solution was to construct two broader entry groups: those who entered before 2.48pm via a turnstile, which he categorised as 'early', and those who entered at 2.48pm or after, categorised as 'later', regardless of whether this was via turnstiles or Gate C. 2.5.80 Crucially this 'later' category also included the composite group whose entry time could only be described as 'probably' after 2.30pm. Even without considering the uncertainty of the assessment, the result as shown in Figure 6 was that those known to have entered between 2.30pm and 2.47pm were placed in the 'early' group, whereas those who entered at an indeterminate time after 2.30pm were included in the 'later' group, even though some or all may have entered before 2.47pm.

¶2.5.81 Dr Nicholl justified this muddle by proposing that the age profile of the indeterminate group was similar to that of the 'later' group. In fact, the age profile would have been just as consistent with the age profile of the 2.30pm to 2.47pm group entering through turnstiles and placed in the 'early' group. This is an unsatisfactory basis for analysis.

171

¶Figure 6: Route and time of entry of those who died As Figure 5, showing construction of 'early' and 'later' groups in Nicholl Report.

¶2.5.82 The second problem evident from the data is that the focus on those who entered 'probably after 2.30pm' is crucial to Dr Nicholl's overall finding. There were many comparisons that could have been made between different entry groups and different blood alcohol levels, but the only comparison that suggested any statistical evidence of a relationship was that which required inclusion of the 'probably after 2.30pm' entrants in the 'later' group, and the 2.30pm to 2.47pm entrants in the 'early' group.

¶2.5.83 All other combinations of entry groups gave results that were likely to have arisen by chance variation alone. Specifically, this includes all analyses omitting the 'probably after 2.30pm' group and all analyses comparing entry before 2.30pm with entry after 2.30pm.

¶2.5.84 The third problem that emerges from replicating Dr Nicholl's analysis is that this sole result that could be described as providing any statistical evidence of an effect also depends on comparing all who had a blood alcohol level of 10mg/100ml or greater with those whose blood alcohol level was reported either as nil or as less than 10mg/100ml. There are clear biochemical reasons, established in the independent Jones Report already discussed, to consider that levels between 10mg/100ml and 20mg/100ml should also be treated as nil, and that levels between 20mg/100ml and 50mg/100ml are either artefactual or insignificant.

¶2.5.85 However, all cut-offs higher than 10mg/100ml, even with Dr Nicholl's flawed construction of 'early' and 'later' groups, produce results that are likely to have arisen through chance alone, and do not provide any evidence of a relationship between entry time and blood alcohol level. It is noteworthy that this includes a cut-off of 80mg/100ml, which was the basis of all of the evidence pursued by the Coroner at the inquests.

172

¶The odds ratio 2.5.86 The fourth problem with the results follows from the use for all Dr Nicholl's comparisons of a statistic known as the odds ratio, an approach followed initially in replicating the analysis of the original data. The odds ratio is generally straightforward to calculate, and some types of epidemiological study cannot generate any better estimate.

¶2.5.87 However, the odds ratio is not an intuitively obvious concept, and often it is erroneously assumed to be the same as relative probability (or relative risk in epidemiological terms). In fact, the odds ratio provides an approximate estimate of relative probability at very low levels of frequency, such as the occurrence of uncommon diseases, but for more common events such as those in this data set the odds ratio differs greatly from relative risk.

¶2.5.88 Dr Nicholl used the correct definition of an odds ratio in his report and in his evidence to the inquests, but at no stage did he clarify that an odds ratio does not estimate the relative probability of the outcome in two different groups in these circumstances, which is what would understandably be assumed by a non-specialist. For example, the only explanation during his evidence to the inquests was:

¶Now one convenient way of expressing this is to say that amongst those victims, the odds of having a raised blood alcohol level for later entrants were three times as great as the odds for earlier entrants. Anybody who is betting on the Cheltenham Gold Cup this after [sic] will understand that as being a useful way of representing this.35

¶2.5.89 Leaving aside the questionable taste of the reference given the circumstances, his first sentence is technically correct (given the flawed definition of entry groups and the inappropriate use of a 10mg/100ml cut-off) but his second is highly questionable.

¶2.5.90 Very experienced punters – or statisticians – may know that the bookmakers' estimate (ignoring their inbuilt 'margin') of the probability of a horse winning that is quoted at evens is twice that of a horse quoted at 3 to 1 against, not three times (odds ratio 3.0, relative probability 0.5/0.25=2.0), but it is unlikely that anybody else will identify the implied exaggeration of the effect. The odds ratio Dr Nicholl quoted for the single statistically significant effect was 3.1 yet the relative probability, which in this case can be calculated from the same data, is less than 2.0.

¶2.5.91 The fifth problem that emerges from replicating the analysis is that in seeking to attribute robustness to his single statistically significant finding, Dr Nicholl crucially misrepresented some results. First, he attempted to counter the criticism that a high proportion of the females and young males amongst those who died both entered the ground before 2.30pm and had low or zero blood alcohol readings, possibly explaining any apparent relationship between time of entry and blood alcohol level.

¶2.5.92 He did this by omitting females and males aged less than 18 years and recalculating the odds ratio relating 'early' and 'later' groups with blood alcohol levels less than 10mg/100ml and 10mg/100ml and greater. He quoted the resulting odds ratio as '3.0 (95% CI: 1.0, 9.3)36 ... exactly as before'.37

Evidence of Dr Nicholl at the Hillsborough Inquest, 14 March 1990, SYC000109160001, pp9-10. 36. 95% CI: confidence interval within which true result is estimated to be with 95% probability given the observed results.

¶undated, SYC000000960001, p23.

173

¶2.5.93 An odds ratio with a 95 per cent confidence interval that includes 1.0 (that is, no difference between the groups) implies that the result was sufficiently likely to have arisen by chance that the finding has borderline significance at best. Hence this cannot be represented as 'exactly as before', where the 95 per cent confidence interval did not include 1.0.

¶2.5.94 More seriously, it is clear from the data that Dr Nicholl resorted to a numerical device to present the lower 95 per cent confidence limit even as 1.0. Recalculating his analysis shows that the true value is 0.97 to two significant figures, so the confidence interval clearly includes unity and provides no statistical evidence of a relationship, but he chose to round to one decimal place instead of two significant figures. This was a dubious approach even in 1989.

¶2.5.95 Dr Nicholl anticipated potential criticism concerning his handling of the 'probably after 2.30pm' group of entrants by reclassifying entrants into two groups, entry pre-2.30pm and entry at or after 2.30pm. The problem, however, remains. He described the odds ratio in this case as '2.9 (95% CI: 1.0, 8.6)', which he interprets in his report as 'some evidence that late entrants after 2:30 were more likely to have raised alcohol levels'.38

¶2.5.96 Again, the lower confidence limit is 0.97, which Dr Nicholl chose to round to one decimal place rather than two significant figures. A 95 per cent confidence interval which extends below 1.0 fails to provide evidence of an effect other than chance.

¶2.5.97 The sixth problem concerns the size of the supposed difference in blood alcohol levels, which Dr Nicholl failed to consider. A small difference can be statistically significant if based on large numbers of observations, but it is unlikely to have any practical importance (for example, a dietary regime that produced a weight loss of 10 grams).

¶2.5.98 The first indication that any difference in this case could only be small comes from the lack of significance in any other comparisons based at higher cut-off values for blood alcohol levels, regardless of how the entry groups are constructed (including Dr Nicholl's flawed construction). The second indication is the small number of people with raised levels of blood alcohol in comparison to the much greater number without, as shown in Figure 7.

¶2.5.99 It is possible to estimate the size of the supposed difference directly, however, although this is not straightforward because of the skewed nature of the data, with all groups (bar one small sub-group of three people) showing a substantial proportion of zero readings. An approach based on regression analysis, for example, is inappropriate as the residual values are non-normally distributed.

¶undated, SYC000000960001, p23.

174

¶Figure 7: Route and time of entry of those who died As Figure 6, showing blood alcohol readings.

¶2.5.100 Yet an approach based on the median difference in blood alcohol level between an 'early' and a 'later' group is possible, defining these groups in the same (flawed) way as Dr Nicholl so as to generate an estimate as favourable to his case as possible. Even under these extreme assumptions, the estimated median difference in blood alcohol level between the two groups is 13mg/100ml (for reference only, approximately 16 per cent of the legal driving limit) and the underlying median difference is unlikely to exceed 38mg/100ml39 (for reference only, less than half the legal driving limit).

¶Weight placed on blood alcohol levels 2.5.101 The Panel's analysis of the original data represents significant criticism of Dr Nicholl's findings and his report. It also brings into question the reliability of his evidence to the inquests, based on his report. At the conclusion of the generic stage of the inquests the Coroner's summing up relied heavily on the Nicholl Report in interpreting the significance of alcohol.40

¶2.5.102 The Coroner stated: 'Of the later entrants, the 2.30 pluses, 43 had had nothing to drink or negligible amounts – I call that nothing – and 22 per cent were over 80 milligrams'. This, he deduced, amounted to 'a fifth in round terms of the people who were those who died who had more than 80'.

Ninety-five per cent confidence interval 0mg/100ml to 38mg/100ml. 40. Dr Popper's summing up of the Hillsborough Inquest, 22 March 1991, SYC00010922, pp46-56.

175

¶2.5.103 In fact this was incorrect, because 16 per cent of those who died had blood alcohol levels above the irrelevant 80mg/100ml marker. Even when referring to later entrants, the 22 per cent figure applied to those in Dr Nicholl's artificially created and paradoxical 'later entrants' group, which did not differ significantly from 'earlier entrants' at the 80mg/100ml level. Thus the correct estimate is 16 per cent.

¶2.5.104 The misleading 22 per cent figure was relied on as significant corroborative evidence by the Coroner:

¶It is very interesting because you may recollect Superintendent Mackay [sic], he was the Detective Superintendent who was standing somewhere around, and he was asked about alcohol and people drinking and he gave an estimate. He said 'I thought about a fifth', about a thousand I think he said people, 'had had perhaps a little bit too much to drink'. That is remarkably close to Dr Nicholl's figures ... That I thought was quite interesting.

¶2.5.105 The Coroner further developed this aspect:

¶... a fifth would probably be about right and that is in fact, as I have told you, what

¶Superintendent Mackay estimated, which I think was very smart of him. It was also very smart of Mr Creaser because he had described the people whom he saw as '3 pint men' which fits in exceedingly well [with] what I have told you the people selling the drinks told you, and it also fits in exceedingly well with Dr. Nicholson's [sic] figures, if you think about [sic].

¶2.5.106 This mix of unreliable 'scientific' evidence and unsubstantiated opinion underpinned the Coroner's summing up to the jury on the possible effect of alcohol:

¶[L]et's ... say 20% and let's take it that they were the '3 pint people'. What effect if any did that have on their behaviour and in particular their response or otherwise to direction and on their mood in the sense of increasing their frustration; decreasing their frustration; increasing their aggressiveness; decreasing their aggressiveness, or what effect did it have ? ...

¶Mr Marshall was quite clear. He did not say the whole lot were drunk. On the contrary, he said there was a minority, a significant minority he said, but a minority who were affected by alcohol ...

¶The big problem is what effect, if any, can that minority have on the group? We have all had the experience that if you get one person in a group who is loud or misbehaves or does something, that one person can actually cause a disproportionate amount of disruption and that is the problem. You may well find you get caught up in a situation which may not necessarily be of your making but which you cannot do anything about because of the activities of various other people.

176

¶2.5.107 This leaves a clear impression that alcohol consumption was of major significance, particularly when expressed negatively in terms of 'frustration', 'aggressiveness' and 'big problem'. In fact, the literal meaning of the passage would be consistent if these words were replaced by 'well-being', 'calmness' and 'difficult question', yet the impression given would be very different.

¶2.5.108 In contrast to the picture presented to the jury, there was no reliable evidence of a significant link between time of entry to the ground and blood alcohol level among those who died at Hillsborough. Even accepting the flawed foundation on which the original analysis was based, it would show a small and inconsequential difference between earlier and later entrants, unimportant for any practical purpose.

¶2.5.109 Nor was any credible evidence presented that established the relevance of the 'driving limit' threshold. Fans were attending a social function, not requiring the swift reactions and anticipation necessary to control a motor vehicle. Such an inappropriate portrayal, and all that emanated from it, was insufficient to support a reasoned proposition that alcohol played a part in the genesis of the disaster. There was no evidence on which to base the inflammatory rumours, told to the Prime Minister on the day after the disaster, that a 'tanked up mob' charged into the central pens.

¶2.5.110 Finally, the Coroner's interweaving of flawed statistical analysis of the blood alcohol levels of those who died and senior officers' uncorroborated evidence provided a profoundly unreliable indication to the jury that alcohol consumption was a significant element in explaining how the disaster came about.

¶Criminal record checks on the deceased 2.5.111 A solicitor involved in the Hillsborough inquests disclosed a document to the Panel showing that criminal record checks were conducted selectively on some of the deceased who had recorded blood alcohol levels. To protect the privacy of the deceased the Panel has decided not to make public the document but to describe the process through which an attempt was made to establish links between blood alcohol levels and previous criminal convictions.

¶2.5.112 The document indicates that a Police National Computer (PNC) check was conducted on all who died at Hillsborough for whom a blood alcohol reading above zero was recorded. It includes a handwritten list of the names, dates of birth, blood alcohol readings and home addresses of 51 of the deceased and provides screen-prints apparently drawn from the PNC. A summary of the results appears on the front page, establishing the number 'with cons' (convictions).

¶2.5.113 The document was not formally part of the West Midlands or South Yorkshire Police inquiries and there is no record in the documents provided by either force or by the Coroner. There is no record of who conducted the checks or precisely when the checks occurred. The National Policing Improvement Agency, the organisation responsible for the PNC, confirmed to the Panel that information has not been retained within the PNC.

¶2.5.114 It is the Panel's view that criminal record checks were carried out on those of the deceased with recorded blood alcohol levels in an attempt to impugn personal reputations. There is, however, no evidence to suggest that this inappropriate – and possibly unlawful – exercise was used in the investigations, inquiries or inquests.

177

¶Blood alcohol levels in survivors 2.5.115 It was known that blood alcohol levels were tested in those who died, because of the prominence given to the results during the mini-inquests. It has not been previously recognised that blood alcohol levels were tested in at least some of the survivors, but this is the implication of some of the material disclosed to the Panel.

¶2.5.116 The most clear-cut evidence is a document among medical papers headed 'In strict confidence',41 continuing 'Blood Alcohol concentrations in samples taken from patients admitted to the Royal Hallamshire Hospital following the Hillsborough Disaster'. There follows a list of 11 names (redacted as confidential medical information) and/or 'Majax Numbers '42 and the corresponding blood alcohol levels, which were all 'not detected' bar two.

¶2.5.117 The same set of documents also contains some text apparently intended to be put onto 'acetates' for overhead projection.43 Under the heading 'ALCOHOL', the text notes the numbers of deceased with alcohol levels of over 80mg/100ml (15) and over 120mg/100ml (6). The text continues: 'FEW OF THOSE ADMITTED HAD APPRECIABLE LEVELS'.

¶2.5.118 It is clear from these disclosed documents that blood alcohol levels were tested in some of those taken to the Sheffield hospitals. Two questions arise: for what reason were these tests carried out, and how extensive was the testing?

¶2.5.119 The individual hospital notes disclosed to the Panel are not of direct help. The only notes that contain reference to blood alcohol are those of a person who survived for two days before dying. Both the laboratory report, naming the pathologist who conducted the post mortem, and the relevant preliminary hearing transcript ('Yes, blood alcohol, this was done on a specimen taken at the time the patient was admitted'44) suggest that this test was carried out after death on a blood sample taken for another purpose on admission.

¶2.5.120 No other medical notes that were traced contained reference to blood alcohol testing, or any reference to the results, including the notes of those identified in the list of 'Blood Alcohol concentrations in samples taken from patients admitted to the Royal Hallamshire Hospital following the Hillsborough Disaster'.

¶2.5.121 The absence of reference to blood alcohol testing in the medical notes does not help to clarify how extensively this testing was carried out, but it is of concern. If these tests were done as part of clinical care – for example to indicate whether alcohol consumption might have contributed to reduced consciousness levels – the results should have been filed in the notes.

¶2.5.122 Further, the notes of some of those identified in the Royal Hallamshire Hospital list, where available, show no medical reason to test blood alcohol levels. The list includes individuals who were fully conscious and orientated, were suffering only from minor injuries, and were not admitted to hospital.

¶2.5.123 The Panel was concerned to trace all relevant documents that might explain why blood alcohol levels were taken and in how many people. No further information has been disclosed but the decision remains contentious and disturbing. 41. File labelled 'AC Crosby', undated, JWR000000250001, p1. 42. 'Majax Numbers': consecutive identifying numbers given to casualties resulting from a major incident on arrival at hospital, pending subsequent confirmation of identity. 43. File labelled 'Hillsborough Reports', undated, JWR000000220001, p1. 44. Inquest transcript, 1 May 1990, SYC000109960001, p8.

178

In the great majority of cases, the cause of death given after post mortem examination was either traumatic asphyxia or crush asphyxia, each regarded as synonymous terms. The disclosed documents show that this corresponded to an assumption made by the Coroner and formed before the post mortems were conducted.
The detailed review of all post mortem reports casts significant doubt on the single unvarying pattern, described consistently during the 'mini-inquests', of traumatic asphyxia causing unconsciousness within seconds, followed inevitably by death within a few minutes.
There was clear evidence from the post mortem reports that 28 of those who died did not have traumatic asphyxia with obstruction of the blood circulation, and asphyxia may have taken significantly longer to be fatal. There was separate evidence that in 31 the heart and lungs had continued to function after the crush, and in 16 of these this was for a prolonged period. (These numbers cannot be added to the 28 as some featured in both groups.)
It was asserted repeatedly, by the Coroner, by the High Court in the Judicial Review proceedings and by the Stuart-Smith Scrutiny, that the effects of asphyxia were irreversible by the time each of those who died was removed from the pens. Yet individuals in each of the groups now identified could have had potentially reversible asphyxia. Resuscitation of an unconscious person with a heartbeat is much more likely to be successful than if cardiac arrest has already occurred, as was previously assumed. While they remained unconscious, these individuals were vulnerable to a new event, particularly further airway obstruction from inappropriate positioning.
It is not possible to establish with certainty that any one individual would or could have survived under different circumstances. It is clear, however, that some people who were partially asphyxiated survived, while others did not. It is highly likely that what happened to these individuals after 3.15pm was significant in determining that outcome. On the basis of this disclosed evidence, it cannot be concluded that life or death was inevitably determined by events prior to 3.15pm, or that no new fatal event could have occurred after that time.
Disclosed documents provide no rationale for the Coroner's exceptional decision to take samples for blood alcohol measurement from all of the deceased.
The implicit and explicit use of a blood alcohol level of 80mg/100ml as a marker was unjustified. This level has relevance to the rapid response times of individuals in charge of motor vehicles, but none to people attending a leisure event.
Analysis of the data demonstrates that the attempt to draw statistical correlation between the time of arrival and alcohol level was fundamentally flawed in six respects, and no such link could be deduced.
The weight placed on alcohol levels, particularly in the Coroner's summing up at the inquests, was inappropriate and misleading. The pattern of alcohol consumption among those who died was unremarkable and not exceptional for a social or leisure occasion.
A document disclosed to the Panel has revealed that an attempt was made to impugn the reputations of the deceased by carrying out Police National Computer checks on those with a non-zero alcohol level.

179

The disclosed documents show that blood alcohol levels were tested in some survivors who attended hospital, as well as in all those who died. There is no record of these tests or their results in the medical notes of survivors, and in some there was no apparent medical reason for the test. The extent of this testing remains unknown.
There was no evidence to support the proposition that alcohol played any part in the genesis of the disaster and it is regrettable that those in positions of responsibility created and promoted a portrayal of drunkenness as contributing to the occurrence of the disaster and the ensuing loss of life without substantiating evidence.