Admitted failings in Article 2 inquests needn’t be part of the Coroner’s conclusion

R (Carole Smith) v HM Assistant Coroner for North West Wales [2020] EWHC 781 (Admin) 7.4.2020

The High Court has emphatically supported a Coroner’s decision not to record admitted, non-causative failings in an Article 2 inquest conclusion. The court’s observations about the proper content of a Record of Inquest (ROI) will be thought-provoking for all Coroners and practitioners particularly as they (i) imply judgmental conclusions on the ROI may be inappropriate (ii) narrow the import of the Tainton decision to jury cases and (iii) elevate the Coroner’s findings of fact and pre-conclusion reasoning to a level that may have significant repercussions for how Coroners close their proceedings in future.

The facts

Leah Smith was just 27 years old when she hanged herself at home, a month after her first presentation with paranoid delusions. During that time she had numerous engagements with a home treatment team, and was admitted to hospital following an overdose to be discharged two days later. She was only once seen by a psychiatrist, more than a week after her overdose. After her death, the local health board’s serious incident review recorded a litany of failings: she never received a medical review or a formal diagnosis; the use of anti-psychotic medication was “overcautious”; there was a delay in prescribing antidepressants. The root cause of these problems was “inadequate medical cover for home treatment team patients.”

In the light of that report the Coroner found that Article 2 ECHR was engaged and instructed an expert forensic psychiatrist. The expert’s report was trenchant: there was a “failure of provision of basic medical care…in my opinion, on the balance of probabilities, the death of Miss Leah Smith was not only predictable but was entirely preventable.” But for the health board’s failures, “it is likely that she would have made a good recovery”.

The inquest conclusion

Fast-forward to the inquest conclusion with the Coroner’s findings distilled into a Box 3 that simply recorded that Leah had been found hanging having “had a short history of mental health issues with an attempted overdose a week prior to her death. She was receiving antipsychotic medication and was under the care of the Mental Health Services at the time of her death.” The Coroner then gave a brief, neutral and factual narrative conclusion which did not refer to her clinical care, only noting that “she was suffering from an episode of psychosis of unknown origin”. There were no critical or judgmental findings in Box 3 or Box 4 whatsoever, let alone the neglect rider contended for by Leah’s bereaved family.

One can easily imagine this would have surprised Leah’s family, surely advised by their legal team that a highly critical narrative conclusion was all but inevitable given the expert’s uncompromising view. They brought a judicial review against the Coroner, alleging that she had erred in law in the threshold and standard of proof for causation in death, that she had irrationally failed to accept the psychiatric expert’s evidence, that the ROI was not Article 2 compliant, and that she irrationally failed to find neglect. The relief claimed was the replacement of all or part of Box 3 and Box 4 with “a narrative that refers to the failings of care provided by (the health board) to Ms Smith”. Griffiths J glossed this in his judgment as “somewhat imprecise”, but that perhaps reflects the family’s understandable assumption that the outcome of the inquest ought to have been a fait accompli in the light of the criticisms made by the expert and accepted by the Health Board. Could a Coroner be acting properly if their conclusion did not adopt and reflect those criticisms?

The High Court emphatically said YES. It rejected all grounds of judicial review, the neglect ground even being considered unarguable. The reasons given by Dingemans LJ, Griffiths J and the Chief Coroner constitute a primer for all inquest practitioners in some often-misunderstood fundamentals of coronial law, but also point to some of the tensions raised by the High Court’s own recent decision-making.

The key to much of the Court’s thinking was the content of the reasons publicly promulgated by the Coroner before delivering the formal conclusion, and this judgment contains thought-provoking material about the legal and practical status of those reasons.

Statistics alone do not prove causation

The High Court reiterated that the test of causation of death in an inquest is, as stated in Tainton, whether the evidence on the balance of probabilities[1] shows that the conduct in question more than minimally, negligibly or trivially contributed to the death. In considering whether the expert’s evidence was sufficient to show that it was more likely than not that the health board’s failings had contributed to the death, the Coroner cited Chidlow [here]: “The court must look at the Claimant’s individual circumstances, rather than at the general statistics.” In Chidlow, the court held that even when a raw survival rate for the general population is over 50%, the deceased cannot safely be placed in a category of survivors without evidence actually supporting that proposition in their individual case. Being a figure in a statistic does not of itself prove causation.

When the expert’s evidence was tested in court, it became clear his evidence on causation was based in Leah being a figure in a statistic. She had a treatable condition and “in the vast majority of cases (of psychotic depression) patients make a good recovery…over 99% do not go on to kill themselves in the coming few years” he said. But the use of percentage figures like these should not mislead inquest practitioners (or Coroners) into thinking this evidence equates to the balance of probabilities. What causation at an inquest requires is positive evidence on which a finding could be made that the deceased would have fallen into the 99% group and not that rare but indelible 1%[2].

Coroners are not bound to accept expert evidence

Next, Leah’s family made the following, bold proposition to the High Court: the Coroner’s conclusion was irrational because she failed to accept the expert evidence heard about causation. This was in a context where the Health Board “accepted” the pejorative conclusions. The court unsurprisingly responded that as an inquest is an inquisitorial process, the Coroner was not bound to accept her own expert’s evidence. Moreover, the manner in which his evidence was tested in court showed some of his evidence was “little more than an assertion”, with conclusions proffered “on a relatively insecure evidential foundation”. That the health board institutionally “accepted” his evidence was neither here nor there. The Coroner was perfectly entitled to prefer evidence from the treating clinicians which undermined that of her instructed expert and which demonstrated that his analysis had been swayed by “hindsight bias” – something which he had candidly accepted himself.

This should serve as a reminder to Interested Persons of all stripes that in an inquest there are no “parties” determining what is and is not in dispute, and it is often the case, as the prudent lawyer might warn their client, that all bets are off once live evidence begins.

An Article 2 compliant inquest does not have to record ‘failings’ in the conclusion if they feature in the Coroner’s findings… and should not either?

Perhaps the most thought-provoking aspect of the judicial review concerned the family’s argument that the inquest failed to comply with the Coroner’s Article 2 procedural duty, because criticisms of public authorities – which were featured and accepted in the Coroner’s reasons – were not included in the formal Record of Inquest. Counsel for the family produced a draft alternative narrative conclusion, based in what the Coroner had announced as her factual findings, stating inter alia:

“…(Leah) received inadequate care, below the level of basic medical care that a patient can expect to receive from a modern mental health service. Despite an urgent referral, she received no in-person consultation from a psychiatrist until 25 April. In the absence of such consultation, there was no opportunity to reach a proper diagnosis despite florid psychotic symptomatology, suggestive of psychotic depression. Medication (both anti-psychotic and antidepressant) given during much of this time was at a subtherapeutic dose, which risked side-effects. Furthermore, there was no appropriate monitoring of her medication. There were multiple opportunities prior to 25 March, for consultant psychiatrists to have seen Ms. Smith, and no adequate reason for this not to have occurred…”

Strikingly, a Divisional Court including the Chief Coroner found this draft completely inappropriate: “It reads more like a statement of case than the conclusion of a Coroner’s inquest. We cannot approve language of this sort for either Part 3 or Part 4 of the Record.” Putting these matters on the record “would have been wrong…they would have compromised the essential brevity and simplicity required of a conclusion”.

Readers of this blog will now be darting back to look over that language proposed by counsel for the family. That’s right: it is just the kind of thing regularly handed down in Article 2 (and some non-Article 2) inquest conclusions by plenty of Coroners. The High Court has now said very clearly that such language may properly feature in the Coroner’s reasons but does not belong in a narrative conclusion. This is because “as the Court of Appeal said in Jamieson…a verdict (now a conclusion) must be factual, expressing no judgment or opinion…” The court went on: “what Middleton envisages is conclusions of fact as opposed to expressions of opinion…

It is worth remembering that whilst that oft-cited passage of the Chief Coroner’s Guidance No. 17 tells Coroners that judgmental language may be used in an Article 2 narrative, it clearly reminds Coroners and IPs that what Middleton permits is “a judgmental conclusion of a factual nature, directly relating to the circumstances of the death”, and that all narrative conclusions should be kept brief.

The context for these observations was to reject a submission that criticisms ought to feature in a conclusion. It will be fascinating to see how this judgment is now cited by “public body” Interested Persons, both in closing submissions to Coroners and perhaps in due course in the High Court, to argue that judgmental and critical language should be excluded from an inquest conclusion.

Hang on – didn’t Tainton tell us that admitted failings had to be in the conclusion?

It is interesting to note that while Tainton is cited approvingly by the court in the context of causation, in the latter part of the High Court’s judgment, the case goes unmentioned. Did the High Court take it as read that different considerations apply when a Coroner sits without a jury?

It seems so, as the Court said this:

“Both the Reasons and the Record were delivered in public. Both, therefore, were part of the public record. The argument that more of what appeared in the Reasons should have been repeated in the Record has the appearance of an argument of form over substance and we would reject it on that ground alone.”

Although not referred to in the judgment, there are echoes here of the High Court’s ruling in Worthington – featuring the present Chief Coroner – where the validity of drawing a substantial distinction between a Coroner’s findings and conclusion was explicitly rejected.

Whilst the judgment is on first blush inconsistent with the decision in Tainton, where Leveson P and Kerr J found that without admitted non-causative failings – publicly aired at the inquest – forming part of a narrative in the formal ROI, the conclusion would be “materially incomplete and verge on misleading by omission” it must of course be remembered that Tainton was a jury inquest where no findings of fact could be recorded other than those appearing on the ROI. Where a Coroner sits alone, and so can hand down a summary of their reasons, it now appears that admitted non-causative failings need not also be included on the ROI for the inquest to be Article 2 compliant.

Are a Coroner’s “findings” really on the record?

A final thought: is the premise that the Coroner’s reasons are “part of the public record” really correct, on the ground as well as in the imagination of the Divisional Court? The Chief Coroner’s guidance No. 17 directs that, where Coroners sit without a jury, “the key findings of fact should be stated orally in open court, preferably (during or) after the evidence has been summarised (but not written on the record of inquest).

Readers will know well that practice varies significantly between different Coroners, and from inquest to inquest. It is up to Coroners to decide what might constitute “key” findings, and the length and detail of reasons given differs enormously between courts. Although transcripts of audio can be created, it is the Record of Inquest that is, well, the record of the inquest. If not preserved on the ROI, have the Coroner’s findings really formed part of the public record?

In Worthington, Leslie Thomas QC submitted that striking out references to his client’s crimes against his daughter from the ROI would be of benefit to him even though they remained in the Coroner’s written findings, because only the ROI would be “retained and formally referred to”. The court rejected that premise, but they did so in a high-profile, widely reported case where the Coroner had delivered a “very substantial document of nearly a hundred pages, in which the Coroner set out in detail the significant lay and expert evidence and his factual findings…” Is Mr Thomas’ premise not correct in the vast majority of inquests? If so, is there not a risk that directing Coroners to record critical findings in their “reasons” rather than the “record” might hide rather than expose culpable and discreditable conduct?

If a Coroner fulfils the state’s Article 2 procedural duty through their reasoning, and not just their formal conclusion, is the lack of consistency in local and individual practice a sustainable feature of our inquest system? And if the High Court (including the Chief Coroner) think that critical findings should feature in a Coroner’s reasons instead of their formal conclusion, does not Article 2 – and basic principles of open justice – require that those full reasons are properly and consistently promulgated and published?

A new version of guidance No. 17 will surely be following the Supreme Court’s forthcoming settlement of the standard of proof issues raised by Maughan. Watch this space for whether some of the issues raised by this case may also prompt some rethinking of how Coroners go about their most fundamental task of delivering their conclusions.



[1] As a side note: Pedantic inquest lawyers can be forgiven the odd wince when some Coroners, delivering their reasons (often extempore after a long day’s evidence), stray into language that seems to apply the standard of proof somewhat loosely, as in this case, where the Coroner’s summing-up included repeated references to the evidence not providing “certainty”. The High Court agreed that “some passages of (the Coroner’s) reasons, taken in isolation, suggest that the Coroner was applying a test of “certainty”“. The court were not however persuaded that it was “fair or correct to pick out isolated phrases…” The Admin Court will not hold it against a Coroner that they have used such language during a “relatively general and reflective discussion of the evidence”, so long as it can be seen they have ultimately formulated any questions that need to be answered on the balance of probabilities.

[2] At time of writing in April 2020, it is far too early to predict whether many Coroners may face in the near future a number of inquests where it is contended that alleged failings by health professionals or organisations – or even by state authorities more broadly – contributed to avoidable deaths from Covid-19. If they do, statistical analysis about survival rates amongst different population groups is likely to be put before them, but the analysis in Chidlow, applied here, must not be forgotten: general statistics alone do not prove causation.