In this guest blog, Nadine Dougall and Jan Savinc from Edinburgh Napier University, discuss their recent research project to address the evidence gap about linked information to better support clinical judgement and prevent young person suicide.
Content warning – hypothetical scenario involving overdose
Childhood adversity and mental health admissions to hospital prior to suicide (CHASE study)
Imagine a hypothetical scenario, in which a 19 year old young woman – ‘Laila’ – is taken to A&E, having taken too many paracetamol. Laila is crying and distressed, but strenuously denies an intentional overdose, saying it was a mistake. Her Mum is present and says she believes her. A&E staff are unsure and refer onto Psychiatric Liaison, who have no real way of knowing either. Laila is not forthcoming about her life history, and they have not met her before – she had no previous admissions to that A&E and denies any previous overdose attempts. Laila does not have serious physical damage, and so is recorded as ‘accidental poisoning’ and discharged the next day.
In fact, Laila had experienced multiple difficulties and trauma in her life. She had been taken into care aged 13 following the death of her biological mother and physical abuse by her stepfather. While in care, Laila developed anxiety and depression, and was referred by social services to mental health services. During this time, Laila overdosed on paracetamol, was treated, and admitted to a psychiatric facility, where she stayed for several weeks. Laila was then placed with her adoptive Mum in a different area 230 miles away where she did not know anyone. Recently she had started dating her first serious boyfriend, but the relationship had deteriorated. Her partner had resorted to blackmail and Laila had been experiencing escalating anxiety and panic attacks.
Now imagine most of the above information was available to A&E staff, would a different assessment of ‘accidental poisoning’ have been made? What would the outcome have been had there been no gaps in information sharing due to geography and communication gaps between different services 230 miles away? Even if this information had been known, would it have made a difference? What combination of events in Laila’s and other young peoples’ lives are most likely to increase vulnerability to suicide, and therefore should serve as red flags for healthcare staff to use alongside clinical judgement? This was unknown.
Our recent research project sought to address some of this evidence gap.
Suicide is a major cause of death for young people worldwide, with wide-ranging impacts relevant to public health and devastating consequences to families and communities. Suicide rates have been increasing in recent years, with notable increases pre-pandemic in adolescents and young people, therefore there is an urgent need for more suicide prevention. Finding ways of preventing early death by intervening earlier in the lives of children and young people before suicidal behaviours emerge is key.
Published evidence has widely reported risk factors for later suicide including mental health episodes and self-injury. More recently, childhood adversity has also come to the fore, with research studies reporting it as a risk factor for later suicide. What is not known however, is the relative impact of mental health, self-injury, or childhood adversity in any combination along the lifespan before suicidal behaviour develops.
In our research, we aimed to find out when young people who died by suicide had previously been admitted to general or psychiatric hospital for childhood adversity (e.g., maltreatment, neglect, or assault), mental health or self-injury – a time when healthcare practitioners could intervene. We chose a powerful longitudinal design that allowed us to follow the data on health contacts of those who died by suicide and compare them with a control population. This design enabled us to avoid recall bias inherent in many other published studies.
Study data were selected from cradle to grave and this was possible using Scotland’s NHS data and National Records of Scotland death registrations which have been stored routinely since 1981. This enabled us to follow up the health records of 2,477 people born since 1981 who died by suicide up to a maximum age of 36. We were also able to link to people’s maternal health and deaths data. We compared these with 24,777 randomly selected people from the general population matched on age, gender, and geography. We published a study protocol.
We found that 76% of deaths by suicide were young men with an average age of 23. Of the men who died, most (81%) had their first in-patient admissions to a general hospital for ‘assault’ (see ‘main findings 5’ below). Two thirds (68%) of women were first admitted to hospital with co-recorded ‘adverse social circumstances’. We also established that 3.5% females and 2.3% males who died by suicide had experienced maternal death, compared with 0.7% and 1.1% in the general population. Care experience was also a factor, 2.5% of those who died by suicide were admitted from or discharged to care/foster homes or other institutional settings by age 18, compared with 0.2% of the general population.
Regarding mental health, 22% of those who died by suicide (35% women and 18% men) had a first psychiatric diagnosis aged 10-17 compared with only 4% of the general population (6% women and 4% men). Mental health was an umbrella category of a range of conditions, including self-harm. The highest admissions by age 18 for men and women who died by suicide was for ‘self-harm/poisonings’, with 30% of women having had this type of admission, 10 times higher than 3% of women in the general population. Corresponding data for men was 9% and 1%, respectively. In decreasing frequency, remaining admissions were related to alcohol/other conditions, substance use, mood, and anxiety diagnoses. We also found that ‘accidental poisonings’ were more frequent in those who died by suicide, with 3.3% having an admission of this kind compared to 0.4% of the general population.
Excerpts from infographic, view full infographic
However, the key finding of the research was that hospital admissions under age 18 for adversity (e.g., violence or maltreatment) AND a mental health diagnosis (in either order of first admission) produced the highest likelihood of subsequent young person suicide (see ‘main findings 6’). This was higher than mental health admissions only (which subsumed categories of self-injury). More information can be obtained from our research infographic (excerpts above) and the study findings have been submitted for publication.
In conclusion, it is obvious from these research data that our fictitious character Laila was much more vulnerable to suicide given her previous circumstances, some of which may have been recorded in previous healthcare contacts. Healthcare providers should prioritise suicide prevention activity in adolescents admitted as in-patients with previous childhood adversity and mental health records as these were associated with far higher numbers of young person suicide. This study also demonstrated a need for better information sharing between general and psychiatric hospital systems, previously reported elsewhere. Knowing the warning signs to look for with suicidal thinking and behaviour are key, but much more needs to be done around linking services and their data. If we consider that health data systems are not linked up within and across UK nations, it is inevitable that healthcare staff are missing out on information to better support clinical judgement.
The strongest association in young people for a single type of condition was self-injury in adolescence. There is a window of opportunity for healthcare practitioners to identify and flag potential ‘at-risk’ adolescents and provide supportive actions to prevent future suicidal behaviour. Perhaps in future, A&E will have a better ‘red flag’ system to alert staff to previous admissions for childhood adversity and mental health and better support staff with clinical judgements related to potential suicide. Other datasets such as GP data, school, social care, or police records, for example, could be used to extend the search for red flags and assist in getting people the help they need.
Footnote: Both the name Laila and the scenario are fictitious, intended to illustrate a point about the benefits of information sharing.
Nadine Dougall is a Professor of Mental Health & Data Science at Edinburgh Napier University and can be reached at email@example.com or @nadinedougall
Jan Savinc is a Research Fellow in Health Data Science at Edinburgh Napier University and can be reached at firstname.lastname@example.org or @jsavinc