142.In our interim report, we outlined the importance of good quality, accurate data. Poor quality data are undermining the ability to prevent suicide. Access to timely data allows rapid responses (for example, reducing access to places or methods of suicide) to patterns that could indicate suicide clusters, which could help prevent further deaths by suicide. Reliable data will help those evaluating preventative action to know which public health and clinical initiatives are effective in preventing suicide, therefore allowing effective action to be replicated and encouraged.
143.We recognise that ensuring good quality data will not, by itself, save lives. Nevertheless, data is a hugely important issue when seeking to prevent suicide. Bad data is undoubtedly a problem—there is no point looking at and seeking to learn from data if it is not trustworthy. It is difficult to know which public health initiatives are working without reliable data and relying on incorrect data could lead to certain initiatives being advocated when they are actually ineffective.
144.We are disappointed that in the third progress report on the suicide prevention strategy there was no reference made to changing the standard of proof, and the only reference made to coroners was to say that “ONS is also engaging further with coroners to improve the quality of reporting of suicides”. We made clear recommendations in our interim report which would have improved the quality of data including its usefulness for preventing suicide, and it is disappointing that the Government has not included any of them in its latest update report on the strategy.
145.As we explained in our interim report, a conclusion of suicide must meet the ‘criminal’ standard of proof, that is, that the coroner or jury must be certain, beyond reasonable doubt, that the person took their own life and intended to do so. Suicide and unlawful killing are the only two conclusions which must meet this higher standard.
146.The higher standard of proof for suicide is harmful for two reasons. The first is that it increases the stigma around suicide. Ged Flynn, Chief Executive of PAPYRUS, explained why stigma around suicide is dangerous:
The consequences of not being open and acknowledging that the person was instrumental in bringing about their own death is to increase the stigma around suicide. This increases the reluctance of those who are considering ending their lives to acknowledge and speak about their suicidal thoughts. It impedes help-seeking.
147.Professor Louis Appleby, arguing for a change to the standard of proof, told us that
Its equivalence with criminal proof reflects the history of suicide. [ … ] There is a principle here, which is that that standard of proof is a reflection of a system that is full of prejudice and stigma, which we ought to dismantle.
148.The current standard of proof also causes misclassification of deaths by suicide, leading to an underestimation of the numbers of individuals who have taken their lives by suicide. We explain this in detail below.
149.We recognise that different bereaved families and individuals will view the issue of the standard of proof differently. It has been suggested that some families would prefer for the death of a loved one not to be recorded as a suicide, due to the stigma attached to suicide. However we are guided by the evidence of the bereaved families and individuals from whom we heard in the course of this inquiry. PAPYRUS, a majority of whose trustees have lost a child or young sibling to suicide, was clear that you do not deal with stigma by colluding with it:
We understand the reluctance of many parents/partners or family members to hear a suicide conclusion returned following the death of a family member, but the consequences of not being open and acknowledging that the person was instrumental in bringing about their own death is to increase the stigma around suicide.
150.Many of the bereaved families from whom we heard in our inquiry are now actively working to prevent suicide. Our evidence suggests that bereaved families recognise that accurately recording a death by suicide as suicide, though difficult to accept in some circumstances, will achieve a different and far better outcome for future families.
152.If a coroner does not feel that the standard of proof has been met for a death of an individual, he or she has three options. The coroner can declare a conclusion of accidental death; the coroner can declare an open conclusion; or the coroner can choose to solely use a narrative conclusion.
153.Open conclusions are included in the Office for National Statistics’ suicide registrations. However, coroners are discouraged from recording open conclusions. The Chief Coroner’s Guidance No. 17, Conclusions: Short-form and narrative makes clear that
Open conclusions are to be discouraged, save where strictly necessary. An open conclusion should ‘only be used as a last resort, notably when the coroner [or the jury] is simply unable to reach any conclusion on the balance of probabilities as between two competing verdicts’. [ … ]In some cases a narrative conclusion will be preferable to an open conclusion. A narrative will give the coroner (or jury) the opportunity to state what findings are made and what are not.
154.Professor Keith Hawton, Director of the Centre for Suicide Research at the University of Oxford, noted that he and other experts in this field have “considerable concerns about the accuracy of data based on coroners’ verdicts, even when open verdicts are included as possible suicides.”
155.A coroner, when faced with a suspected suicide which does not meet the standard of proof, who is discouraged from using an open conclusion has two options left. The coroner can record the death as accidental (which would not appear in the suicide registrations and would therefore skew the data) or can choose to use a narrative conclusion.
156.Narrative conclusions are not in and of themselves a problem. However, if a coroner has used a narrative conclusion with no short-form conclusion, ONS must code the death. In some circumstances, a coroner will include enough detail about the deceased’s actions and intent for the ONS to code the death as a suicide. However in other circumstances, the coroner will not have included sufficient detail for the ONS to code the death as a suicide. These “hard-to-code” conclusions result in deaths which are likely to be suicides not being coded as such. David Gunnell (a public health physician and researcher and member of the National Suicide Prevention Strategy Advisory Group) notes that in England and Wales, hard to code narrative conclusions are increasing (from 6% in 2011 to 8% in 2014), “compromising suicide prevention activity and leading to significant under-estimation of suicide rates and trends”.
157.Professor Gunnell highlights the variability between coroners and regions:
As the use of narrative verdicts varies tremendously from coroner to coroner over time this may have distorted the assessment of suicide prevention activities in some areas.
158.Research undertaken by Professor Gunnell and other colleagues showed that areas which reported the largest declines in suicide between 2001/2 and 2009 were the areas which experienced the greatest increase in use of narrative verdicts. Some of Professor Gunnell’s research shows huge variation between the decisions of coroners across the country in studies of “clinically defined suicides”. This can lead to ineffective action, as he explains:
I’ve seen claims made for the success of local prevention programmes, where apparent suicide reductions have more likely reflected changes in local Coroner practice (increased use of narrative verdicts).
159.We are concerned that unreliable data could hinder the efforts of public health teams to reduce suicide. We note that if the standard of proof for conclusion of suicide was the civil, rather than the criminal, standard, coroners would be able to record likely suicides as suicide, rather than facing a choice between a conclusion of accidental death, an open conclusion or a narrative conclusion.
160.Apart from a change to the standard of proof, we also consider that improvements to the way in which narrative conclusions are recorded are essential for improving data accuracy. We are encouraged that in a discussion between our Chair and the Chief Coroner on this subject, the Chief Coroner recognised the importance of ensuring consistency between coroners, but he is limited in his powers to change practice due to the independence of coroners and in the resources available to him to help to bring about greater consistency.
161.We recommend that the Chief Coroner should be given adequate resourcing to allow clear oversight of the variation in the recording of suicide. We also recommend mandatory training for all coroners, both those already in post and newly appointed, on the use of short form and narrative conclusions, to ensure consistency across England and Wales.
162.Witnesses have also told us of the importance of inclusion of detail about the intent of the deceased and the method of suicide in narrative conclusions. Detail on intent and method will reduce the number of hard-to-code narrative conclusions, therefore ensuring greater accuracy in ONS data. Systematic and consistent details of method of suicide will also enable quick intervention by public health teams who can act to ensure that the likelihood of further deaths by certain methods is reduced, including restricting access to high-lethality methods of suicide.
163.We recognise the concerns that including detail about lethal (and particularly new or emerging) methods of suicide in coroners’ rulings could be counter-productive, as it could lead to vulnerable individuals using information found in coroners’ conclusions to take their own lives. We suggest that the Government should explore whether information about lethal methods of suicide could be made available to statistical agencies and public health teams, but withheld from public view.
164.We recommend that training for coroners on suicide should include the importance of including sufficient detail in a narrative conclusion about the deceased individual’s intent and method used in order to minimise the number of hard-to-code narrative conclusions. Accurate data is crucial to the understanding of what approaches work best in reducing suicide. We suggest that this training could be given by experts in the field of data and suicide prevention.
165.While acknowledging that coroners are independent judges and that the Chief Coroner cannot direct them to take specific action, we consider that training for coroners could include information about how vital sharing information with public health and mental health teams where appropriate at an early stage can be for reducing the likelihood of a cluster of suicides.
166.We recommend that training and guidance for coroners should include material about the importance of timely information sharing with public health and mental health teams where appropriate in order to identify possible clusters and the proliferation of emerging new methods of suicide.
115 Fourth Report of Session 2016–17, , HC 300, paragraph 27
116 See paragraph 159
117 PAPYRUS (SPR0027), paragraph 11
118 Q296 [Professor Louis Appleby]
119 PAPYRUS (SPR0027), paragraph 11
120 , paragraphs 68, 69 and 73
121 Professor Keith Hawton (SPR0030), paragraph 17
122 Professor David Gunnell (SPR0032), paragraph 7.1
123 Professor David Gunnell (SPR0032), paragraph 7.2
124 Professor David Gunnell (SPR0179)
125 David Gunnell (SPR0179); Samaritans (SPR0176)