G4S Care and Justice Services Ltd v Luke

On behalf of the Defendant, Mr Payne QC submitted that the judge erred in failing to (a) restrict himself to the Claimant’s pleaded case; (b) apply an insufficiently high or stringent threshold in determining whether a ‘real and immediate risk to life existed’; (c) identify any sufficient factors which were capable of establishing that the Defendant should have had knowledge of a real and immediate risk to the life of Mr Boland (as opposed to a generalised unparticularised risk to life) and, a real risk to life as opposed to a real risk of harm; (d) provide adequate reasons for his conclusions.

On the ‘pleaded case’ issue, Mr Payne said that the judge’s finding at [47] that there was ‘the lack of anything which marked Mr Boland out as particularly vulnerable, or any warning signs about the nature of his drug taking in prison’ should have resulted in the judge holding that the claim had no reasonable prospects of success, given the judge did not find that the Defendant’s policies had created a real and immediate risk to Mr

In relation to ‘real and immediate’ risk, Mr Payne pointed out that in this context ‘real’ means ‘significant and substantial’: Rabone, supra, [38]. He said the judge did not apply the right test, and the mere possibility that a prisoner might overdose was not sufficient in the absence of any particular risk to Mr Boland’s life. He said that had the correct test been applied by the judge, he would not have concluded that because drugs were widely available in the Prison there was a real, ie significant and substantial, risk to Mr Boland’s life. Similarly, in relation to the ‘immediate’ limb of the test, Mr Payne said the judge had erred in that there was no evidence of a ‘present and continuing’ risk, which is the meaning of ‘immediate’ in this context: Rabone, supra,

As to the next submission, relating to the particularised risk to Mr Boland’s life from taking drugs, Mr Payne said there was simply no evidence of any such risk because he had promised not to do so and appeared not to be doing so.

Mr Payne also criticised the quality of the judge’s reasons. I say no more about this; it does not appear to me to add materially to the other points which Mr Payne made. This is not a free-standing ‘lack of reasons’ case.

At [37] of his Skeleton Argument Mr Payne went on to make a broader submission to the effect that there can be no operational duty owed to prisoners who have no mental health issues to protect them from the consequences of their own criminal acts, including taking illicit or illegal drugs. Although pleaded in the Defence, this does not appear to have been argued below. He submitted that prisoners who are not mentally ill do not, and cannot as a matter of principle, fall into the recognised categories of persons to whom an operational duty to protect them from risks to their life under Article 2 is capable of being owed. Mr Payne’s essential argument is that although the

On behalf of the Claimant, Ms Hemingway submitted that (a) the judge correctly directed himself on the law and the relevant facts and reached a conclusion that was open to him; (b) in particular, the judge correctly recognised this was a developing area of law which militated against a summary determination of the claim; (c) as an appeal court, I should be slow to interfere with the judge’s conclusion because it was a case management decision; (d) the case raises issues of wider public importance (as the Defendant accepts in [2] of its Skeleton Argument), and this provides an additional reason why it should not be summarily disposed of. She also says that some of the issues argued by Mr Payne went beyond the scope of the single ground of appeal in the Appellant’s Notice and so I should not entertain them: CPR r 52.21(5).

Expanding on these outline submissions, Ms Hemingway submitted that the test I should apply on this appeal is contained in CPR r 52.21(3), and that I can only allow the appeal if the judge’s decision is (a) wrong; or (b) unjust because of a serious procedural or other irregularity in the proceedings below. (I say at once that the second limb has no application on this appeal).

She said I should have considerable deference for the judge’s decision as it was in essence a case management decision, and she relied on Commissioner of Police for the Metropolis v Abdulle [2016] 1 WLR 898, where the defendant appealed against a decision refusing to strike out the claimants’ claims. The Court of Appeal said that the appeal essentially amounted to a perversity challenge and emphasised the need to uphold robust case management decisions of first instance judges.

Ms Hemingway accepted that the test is one of real and immediate risk, and that they bear the meanings advanced by Mr Payne, namely ‘real and substantial’ and ‘present and continuing’.

She submitted that Mr Payne was not correct to say the judge wrongly failed to identify sufficient evidence of a real and immediate risk to Mr Boland’s life individually as opposed to some risk to prisoners in general. She said there were a number of cases from the Strasbourg court where it has been held that the Article 2 protective obligation can be owed to members of the public at large without the need for a specific individual to be identifiable in advance, for example, Mastromatteo v Italy, Application Number 37703/97, where the Grand Chamber was willing to accept the risk posed to the general public by criminals on day release, even though the deceased victims were not identified prior to their encounters with the perpetrators. She also relied on Tagayea v Russia, Application Number 26562/07, where a threat of terrorist attack posed to all school children in multiple districts, even where the individuals had not been identified with precision, was sufficient to impose an operational obligation on the state to put protective measures in place. She also pointed to domestic decisions to the same effect, including Sarjantson v Chief Constable of Humberside [2014] QB 411 and R v Marine A [2014] 1 WLR 3326.

Ms Hemingway argued that the judge had been right to conclude that there was a drugs problem at the prison. She emphasised Lord Dyson’s judgment in Rabone, supra, [35][38], where he had said that ‘real’ did not mean ‘likelihood or fairly high degree of risk’. She said that the question of whether the necessary level of risk was established was very fact sensitive, and that it was impossible to fault the judge’s conclusion that there was a proper basis upon which a trial judge could find that the risk of Mr Boland’s death was real.

As to the ‘immediate’ limb of the test, Ms Hemingway submitted that the judge was right for the reasons he gave at [47] to hold that there was an immediate risk.

As to the Defendant’s knowledge, she submitted that the evidence of prison witnesses who had seen the effects of Buscopan and Black Mamba was sufficient to fix the Defendant with the requisite knowledge.

Ms Hemingway also submitted that the ‘pleaded case’ point was not in the Appellant’s Notice and I should not allow it to be advanced. However, in any event she said there was nothing in it.

Overall, she submitted that the judge reached the right conclusion for the right reasons and that the appeal should therefore be dismissed.

Ms Hemingway also seeks leave to adduce fresh evidence not before the judge below, pursuant to Ladd v Marshall [1954] 1 WLR 1489. The evidence is a witness statement

Ms Hemingway says that I should receive this evidence because the three Ladd v Marshall, supra, criteria are satisfied in that (a) the evidence could not have been obtained with reasonable diligence for use in the court below; (b) it would have an important influence on the result of the case, although it need not be decisive; (c) the evidence must be credible.

The Defendant accepts that the first and third of these criteria are satisfied, but disputes that the second criteria is satisfied. Mr Payne says that the fresh evidence does not advance the Claimant’s case because it does not add anything which is capable of supporting the contention that that the Defendant knew or ought to have known of a real and immediate risk to Mr Boland’s life, for the reasons given above.

I have read this evidence de bene esse and will deal with it later in my judgment.

The test to be applied on this appeal 60. CPR r 52.21(3) provides:

“(3)

The appeal court will allow an appeal where the decision of the lower court was -

(a)

wrong; or

(b)

unjust because of a serious procedural or other irregularity in the proceedings in the lower court.”

As I have said, the second limb is not in play in this case. The Claimant submits that judge’s decision was an exercise of discretion in relation to a case management decision, and thus I should have considerable deference to it. I disagree. The case of Abdulle, supra, which the Claimant relied on, concerned a refusal to strike out a statement of case for failure to comply with a rule, practice direction or court order, which in turn involved an exercise of discretion, as the headnote to the case makes clear. Here, the judge’s task was to decide on the materials before him whether the Defendant had shown that the Claimant’s case had no realistic prospect of success. That was a question to which there was only one correct answer. The issue for me is whether the judge’s conclusion was wrong.

Although this issue formed the second part of Mr Payne’s Skeleton Argument, I consider that, logically, it should be addressed first. That is because, if the operational duty is incapable of applying to a prisoner in respect of such a risk, then it is irrelevant whether or not the risk is real and immediate. This was the approach of the Court of Appeal in Rabone, supra: see at [2011] QB 109, [63], where Jackson LJ said:

“On the law as it stands, I do not believe that health trusts have the article 2 operational obligation to voluntary patients in hospital, who are suffering from physical or mental illness, even where there is a real and immediate risk of death.”

Although the Supreme Court reversed the Court of Appeal and held that the operational duty is owed to voluntary psychiatric patients, the underlying point in this passage is sound (if I may respectfully say so): if a person is not within a category of individual to whom the operational duty can apply then the level of risk does not matter.

I referred towards the start of this judgment to the operational duty being owed to categories of those who can generally be regarded as ‘vulnerable’ individuals. In Rabone, supra, [15]-[16], Lord Dyson set out the categories of such persons as has (at least then) been recognised by the Strasbourg court. He said:

“15.

The court has held that there is a duty on the state to take reasonable steps to protect prisoners from being harmed by others including fellow prisoners (Edwards v UnitedKingdom (2002) 35EHRR487) and from suicide (Keenan v United Kingdom (2001) 33EHRR913). The same duty exists to protect others who are detained by the state, such as immigrants who are kept in administrative detention (Slimani v France (2004) 43EHRR1068) and psychiatric patients who are detained in a public hospital ( Savage v South Essex Partnership NHS Foundation Trust (MIND intervening) [2009] AC 681). The operational duty is also owed to military conscripts. Specifically, there is a duty to protect a conscript against the risk of suicide ( Kilinç v Turkey (Application No 40145/98) (unreported) given 7 June 2005) …

16.

More recently, the court has expanded the circumstances in which the duty is owed so as to include what may generally be described as dangers for which in some way the state is responsible. Thus inÖneryildiz v Turkey41 EHRR 325, the applicant had lived with his family in a slum bordering on a municipal household refuse tip. A methane explosion at the tip resulted in a landslide which engulfed the applicant’s house killing his close relatives. The Grand Chamber held, at para 101, that the Turkish authorities knew or ought to have known that the tip constituted a real and immediate risk to the lives of persons living close to it. They consequently

“had a positive obligation under article 2 of the Convention to take such preventive operational measures as were necessary and sufficient to protect those individuals, especially as they themselves had set up the site and authorised its operation, which gave rise to the risk in question.”

It is clear, however, that the categories of such persons to whom the operational duty may be capable of applying are not closed. As Lord Dyson went on to say at [25]:

“Strasbourg proceeds on a case by case basis. The jurisprudence of the operational duty is young. Its boundaries are still being explored by the ECtHR as new circumstances are presented to it for consideration. But it seems to me that the court has been tending to expand the categories of circumstances in which the operational

duty will be found to exist.”

Also, more recently in Daniel v St George's Healthcare NHS Trust [2016] EWHC 23 (QB), [20] Lang J remarked that ‘the application of the operational duty has developed incrementally’.

The notes to CPR r 24.2.3 in the White Book 2019 make clear that an application for summary judgment is not appropriate to resolve a complex question of law and fact. In his judgment in Rabone, supra, at [22]-[24], Lord Dyson considered the indicia in the Strasbourg jurisprudence as to when an operational duty may be found to exist in relation to a particular category of person. He demonstrated that the issue is heavily fact-specific, depending, among other things, on the type of risk and level of vulnerability of the individual in question.

Thus, the categories of the ‘vulnerable’ are developing, and whether the duty exists in relation to a particular category of person requires a heavily fact specific analysis in light of the underlying principles. Both of these factors mean that in my judgment it would not be right to summarily determine the claim on the broad basis argued for by the Defendant, and I decline to do so. I recognise the force of Mr Payne’s submissions as to why there ought not to be an operational duty to protect a prisoner from risks to his life arising from his voluntary choice to take illicit drugs or from his own criminal conduct. But that broad submission fails to take account of the fact that the approach of the Strasbourg court is always heavily fact specific, and certainly is so in this context. The operational duty applies to prisoners as regards certain types of risk, as the cases show. There might, in a particular case, always be particularly special facts which mean that whatever the general position, in respect of a particular prisoner, such a duty might be capable of applying in relation to the risk arising from his choice to ingest drugs obtained in prison.

I will therefore assume (without deciding) that the operational duty imposed by Article 2 is capable of extending to a prisoner in respect of a risk to his life caused by the voluntary ingestion of drugs which he has illicitly obtained whilst in prison, and that in this case it extended to Mr Boland. The question before me is therefore whether the judge was wrong to hold that there was a realistic prospect of the Claimant showing that there was a real and immediate risk to Mr Boland’s life which the Defendant knew of, or should have known of.

I begin my discussion of this topic with the nature of the Osman ‘real and immediate’ test.

It has been consistently emphasised that this test is not easy to satisfy. In other words, it is a stringent test. In Van Colle, supra, at [115] Lord Brown of Eaton-under-Heywood explained:

“115.

The test set by the European Court of Human Rights in Osman v United Kingdom (1998) 29 EHRR 245 and repeatedly since applied for establishing a violation of the positive obligation arising under article 2 to protect someone from a real and immediate risk to his life is clearly a stringent one which will not easily be satisfied. This is hardly surprising given, as the Osman judgment itself recognises (at para 116), “the difficulties involved in policing modern societies, the unpredictability of human conduct and the operational choices which must be made in terms of priorities and resources”. It is, indeed, some indication of the stringency of the test that even on the comparatively extreme facts of Osman itself—rehearsed by Lord Bingham at para 56—the Strasbourg court found it not to be satisfied.”

Statements at the highest level to the same effect are not hard to find. In Van Colle, supra, [30], Lord Bingham of Cornhill said the test was ‘not easily satisfied’ and that ‘I would for my part accept that a court should not lightly find that a public authority has violated one of an individual’s fundamental rights or freedoms’. In the same case at [69] Lord Hope of Craighead said that the test imposes ‘a very high threshold’. In Re Officer L [2007] 1 WLR 2135, [20] Lord Carswell said that, ‘It is in my opinion clear that the criterion is and should be one that is not readily satisfied: in other words, the threshold is high.’

The Osman test imposes a higher threshold than establishing mere negligence. In Rabone, supra, [36]-[37] Lord Dyson said:

It is also important to emphasise that the risk must be of a risk to life; a risk of serious harm is not sufficient. Kent County Council, supra, is a tragic case where a formerly high-achieving teenage boy, EB, starting missing school and then went into a downward spiral over a number of months which involved a number of arrests for shoplifting, being taken home from a park by the police drunk or drugged, associating with street drinkers and drug takers and increasingly being absent from home overnight, occasionally sleeping rough. There was also a suicide attempt. Eventually, he died aged 14 after drinking methadone. The Council argued that the coroner had been wrong to hold that there should be an Article 2 Middleton inquest: R (Middleton) v West Somerset Coroner [2004] 2 AC 182. It submitted that there was no operational duty to take protective measures because, inter alia, there was no real or immediate risk to EB’s life which the Council knew about or ought to have known about. The Council accepted that it was arguable that there was some risk of potential harm to EB, but that that risk did not extend to a real and immediate risk to his life.

The Court agreed, saying at [44]-[47] that:

“44.

There is, in our judgment, no doubt that the instant case, involving a vulnerable child in the circumstances we have outlined, enters into the potential territory of operational duty. It follows that the question to be considered in the first place, applying the test set out in paragraph 41 above, is whether there was a real and immediate risk to the life of EB in the period before his death.

45.

There is no doubt that the tragic outcome of death came as a shocking surprise. It was undoubtedly true that in the nine months that the claimant knew about EB it knew that his problems had become more serious over that period. There was, as we have said, a downward spiral. There was undoubtedly a need for some action and it could be said that the claimant failed him in this regard …But even if the opportunities had been taken by the claimant, there is no certainty or even likelihood that EB's death would have been avoided as the SCR concluded.

46.

EB was undoubtedly vulnerable and at risk. But that risk, viewed objectively, was not a risk to life, certainly not a real and immediate risk to life. It was a risk of harm, but that should not, with all the wisdom of hindsight, be equated with a risk to life. Ms Gallagher identified that risk of harm in the Executive Summary of her Report, but she did not identify a risk to life. She identified the risks which he faced from his behaviour, but while she identified the risks as involving possible "significant harm", at no stage did she say that the risk should have been assessed as more serious.

47.

It follows, therefore, in our judgment, that there was no sufficient evidence of ‘a real and immediate risk to life’ …”

Although in Osman, supra, the Court referred at [116] to a real and imminent risk to the life of ‘an identified individual or individuals’, that arose out of the particular facts of the case. The subsequent jurisprudence shows that the Court has not limited the scope of the Article 2 duty to circumstances where there is, or ought to be, known a real and imminent risk to the lives of identified or identifiable individuals. It is sufficient if

The Osman test requires the public authority to have known, or that it should have known, of the existence of the risk. As to this, guard must be taken against hindsight: what matters is what the public authority knew or ought to have known at the time. In Van Colle, supra, Lord Bingham said at [32]:

“32 In its formulation of the “real and immediate risk” test the Strasbourg court, in para 116 of its Osman judgment, laid emphasis on what the authorities knew or ought to have known “at the time”. This is a crucial part of the test, since where (as here) a tragic killing has occurred it is all too easy to interpret the events which preceded it in the light of that knowledge and not as they appeared at the time.”

Where there is such a risk, then the standard demanded for the performance of the operational duty is one of reasonableness. This brings in ‘consideration of the circumstances of the case, the ease or difficulty of taking precautions and the resources available’: per Lord Carswell in Re Officer L, supra, [21].

It is now time to apply these principles to the facts of this case and the judge’s analysis.

Earlier in this judgment I set out the relevant part of the Particulars of Claim setting out why the Claimant says there was a real and immediate risk to Mr Boland’s life. I accept that these points are non-exhaustive. But they must provide at least a foundation for the Claimant’s case.

I begin with the issue of the availability of drugs in the Prison. I accept that the judge was entitled to find that illicit drugs such as Black Mamba were easily available on the wing where Mr Boland died. I also accept that the judge was entitled to find that Buscopan (a prescription drug) was being misused by some prisoners to whom it had not been prescribed, and that prison staff had witnessed prisoners undergoing very unpleasant side effects such as hallucinations. I also accept the judge was entitled to find that this fixed the Defendant with the knowledge that such drugs were being abused by some prisoners. However, it is self-evident that misusing any drug carries a risk to health. That is true even of common over-the-counter medications such as paracetamol and ibuprofen. But to conclude from this generalised risk of harm that there was a realistic prospect of the Claimant showing that misuse of these drugs carried a real risk of death paid insufficient regard to the fact that the Osman test is a stringent one which is not easily satisfied. A real risk of harm does not equate to a real risk of death, as the Court reasoned in the Kent County Council case, supra. In short, I agree with Mr Payne that the judge did not apply the Osman test in a way which paid proper regard to its stringent nature. His conclusion at [45] ‘… there is enough to establish that the Defendant knew or ought to have known that the use of Buscopan, particularly with other drugs, might cause death’ was not one which was open to him on the findings that he made or on the evidence before him if by it the judge meant that the Osman test was satisfied. In fact, the evidence was uncertain as to how the drugs in question interacted together. I note the expert toxicology evidence from Rochelle Savage in her statement of 26 June 2015 that ‘The potential interaction between synthetic cannabinoids and hyoscine [ie, Buscopan] when abused with other drugs is unknown.’ Also, the Prison Ombudsman concluded at [55] of his report, ‘We are satisfied that Mr Boland had been warned of the dangers of mixing drugs and the possible lethal consequences.’

There is nothing in the fresh evidence relied upon by the Claimant to alter this conclusion. I readily accept that this shows a bad situation in the Prison so far as the availability of drugs and other contraband items are concerned and that matters have not significantly improved since Mr Boland died. Also, that there were four prisoner deaths in March 2018 is, of course, a matter of real concern. But the evidence does not provide much, if any, further support in showing that there was a real and immediate risk to life from the use of illicit drugs. As far as these death are concerned, perhaps of greater relevance to the question of whether the Defendant knew or ought to have known of a real and immediate risk to Mr Boland is that at the date of the Prison Ombudsman’s report into his death (May 2016), there had been 12 deaths since 2013, including Mr Boland. Five of these were self-inflicted, while the Coroner found that his death had ‘no significant similarities with the circumstances of the other cases.’ Thus, the pattern of prison deaths was not such as to put the Defendant on notice of a risk of death from abusing Buscopan and Black Mamba.

In my judgment, there is another reason why the judge’s conclusion was wrong. Even if the judge had been correct to hold that the use of Buscopan and Black Mamba with other drugs gave rise to a real and immediate risk of death, I agree with Mr Payne that the judge needed to go on to consider on what basis it could be said that there was a real and immediate risk to Mr Boland’s life. It was not sufficient that he was a member of the prison population where such drugs were rife. Even though the judge was entitled to find that misuse of drugs was taking place on B Wing, it does not follow that all prisoners were at risk as a consequence. Some prisoners choose to take drugs illicitly, others do not. The mere fact that drugs were available did not of itself create a real and immediate risk to the life of a given prisoner. Nor was the fact that Mr Boland was a

Ms Hemingway is certainly correct to point out that, as discussed in Sargantson, supra, the Strasbourg Court has not limited the scope of the Article 2 duty to circumstances where there is or ought to be known a real and imminent risk to the lives of identified or identifiable individuals. It may be sufficient if there is such a risk to a group of individuals, which may include society at large. But it has done so in cases such as Öneryildizsupra, where people were at risk because of their status or situation alone (eg as residents living on the edge of a refuse tip): nothing more was required to put them at risk. Mr Boland’s situation was different. He was not at any risk from the drugs in the prison unless and until he decided to ingest them. Hence, in order for the Osman test to be capable of being satisfied, there had to be evidence that the Defendant knew or ought to have known that he was abusing illicit drugs including Buscopan and Black Mamba, the two drugs which the judge focussed on as being principally responsible for the risk of death which the judge found to have existed. As I have said, I consider he was wrong on this issue, but even if he had been right, then what was needed was some evidence that Mr Boland specifically was taking these drugs.

There was no such evidence. The judge found at [45] that there was no direct evidence that the Defendant was taking Black Mamba and Buscopan and other drugs. In light of that finding, the judge should have gone on to find that there was no realistic prospect

In his judgment the judge relied on the Prison Ombudsman’s report into Mr Boland’s death. A number of factual matters in that report are relevant. And, it seems to me, they are not capable of being undermined, whatever may come out by way of further disclosure or evidence. These matters are: (a) that on reception into prison in the days following 29 January 2015 Mr Boland confirmed his drug/alcohol abuse but said that he had not self-harmed for 10 years and had no thoughts of suicide or self-harm; (b) he met with the Prison’s Integrated Drug Treatment Service (run by the NHS, which provided healthcare services to the Prison), who undertook a risk assessment which identified no risk of medication abuse, self-harm or suicide; gave Mr Boland advice as to the dangers of taking illicit drugs in addition to those which he was prescribed; and prescribed him a number of drugs to treat his opiate dependence, alcohol withdrawal and depression; (c) that on 3 February 2015 Mr Boland signed a compact by which he promised, among other things, not to stockpile medication nor have in his possession drugs which had not been prescribed to him; (d) thereafter, that he engaged with medical staff; (e) at the end of March 2015 he told his drug worker that he was ‘drug free’. He also said, according to the Ombudsman, ‘… he felt OK, was doing well with his medication and no thoughts of suicide or self-harm’; (f) of the six drug tests he took in prison he passed five of them. The one he failed was for opiates on 17 February 2015, some two months before his death. He passed the five subsequent tests that he took before his death; (g) between 13 and 15 April Mr Boland attended a ‘motivation to change’ course in which he had played an active part. He told a nurse that he was likely to be released in June and wanted to be stable on his medication.

In light of these incontrovertible matters, and having regard to the stringent nature of the Osman test, it is impossible, it seems to me, to conclude that there was a real and immediate risk to Mr Boland’s life of which the Defendant knew, or ought to have known. I take the point that the drug tests were not capable of testing for Black Mamba. But looking at the picture overall, as far as the Defendant knew, whatever the general prevalence of drugs within the prison, Mr Boland was not at risk because it appeared that he had finally managed successfully not to take drugs for a sustained period and at the time of his death had been drug free for some time.

I have every sympathy for Ms Luke for the death of her son. But, for the reasons that I have given, the judge’s decision was wrong. There is not a realistic prospect of the Claimant showing that there was a real and immediate risk to Mr Boland’s life, of which the Defendant was or should have been aware, so as to trigger its operational Osman duty. This appeal must therefore succeed. There will be summary judgment for the Defendant on the Claimant’s claim.