When I give talks about self-harm and suicide in children and young people, many in the audience are often shocked because (i) I look much older in real life than in my outdated faculty photo (see above) and (ii) because of the surprising statistics I discuss: notably that suicide is the leading cause of death in 5-19 year olds in England and Wales (ONS) and globally the second largest cause (WHO). Moreover, we know that 50-60% of children and adolescents who tragically die by suicide have self-harmed previously (NCISH, 2017) and this increased risk of suicide can last for many years (Hawton et al 2020).
Thus, understanding and responding effectively to self-harmful behaviour is a crucial element of suicide prevention (Townsend, 2019). With all of this in mind, plus the dearth of research on ‘suicidality’ (taken here to mean suicidal thoughts, plans and behaviours) in younger children, this new research (Janiri et al, 2020) is a timely addition to knowledge.
The Adolescent Brain and Cognitive Development (ABCD) cohort study investigates mental health from childhood to adulthood, collecting a comprehensive range of measures including environment (social and family), wellbeing (physical and mental), brain scanning, genetics, cognition and behaviour. The sampling covers 20% of the population of the US in the 9-10-year-old age group.
Established measures are used such as the Child Behavior Checklist (CBCL) and the computerised Kiddie Schedule for Affective Disorders and Schizophrenia for DSM-5 (KSADS-5) which includes a range of items relating to self-destructive thoughts and behaviours.
This study (Janiri et al, 2020) presents cross-sectional data and focuses on thoughts and behaviours that were considered suicidal in nature, and excludes items without suicidal intent. Both caregiver (mostly mothers) and child reports were gathered.
Findings were based on a sample of 7,994 unrelated children with a mean age of 9.9 years of whom 53% were boys. Around eight in every 100 children reported experiencing ‘suicidality’. The headline findings based on regression models (accounting for 24% and 57% of the variance in child-reported and caregiver-reported ‘suicidality’ respectively) were that:
- Regardless of informant, family conflict and childhood psychopathology (as measured by the CBCL total score) were the most robust risk factors for ‘suicidality’ (any suicidal feature either in the past or currently)
- Risk of child-reported ‘suicidality’ increased with higher weekend screen time
- Risk of child-reported ‘suicidality’ decreased with greater parental involvement and positive school involvement
An identical pattern was found in sex-specific analyses (the same for girls and boys) though screen time and parental supervision were ‘… more consistently associated with suicidality in boys than in girls.’ (p.6)
The concordance of caregiver-child reporting on each aspect of ‘suicidality’ was extremely low, with caregivers generally reporting fewer issues (e.g. child-reported suicide attempts totalled 59 compared to 23 caregiver-reported attempts).
In children aged 9-10 years old, childhood psychopathology and family conflict emerged as the most significant risk factors for suicidal thoughts, plans and behaviours. However, the authors note, ‘Even after accounting for psychopathology, children who reported family conflict were 30-75% more likely to experience suicidality.’ (p7) highlighting the pivotal importance of family relationship problems and conflict.
Moreover, eight in every 100 children reported experiencing ‘suicidality’. The authors note that the low concordance between caregiver and child reporting of suicidality means that these factors cannot be reliably assessed by caregiver report alone. Higher weekend screen use (watching TV, using a computer and texting) was also associated with increased ‘suicidality’.
In terms of protective factors, greater parental supervision and positive engagement in school appeared to mitigate ‘suicidal’ risk.
The authors conclude that their study results have:
… immediate and practical implications as risk factors (childhood psychopathology, family conflict and screen use) and protective influences (higher parental supervision and positive school engagement) are actionable and modifiable. (p.9)
Strengths and limitations
A key strength of this study is the large, nationally representative sample which included an almost 50:50 gender balance. This enabled the researchers to investigate a wide range of validated measures robustly and both child and caregiver responses were included.
Few studies have examined protective factors alongside risk factors, and hardly any report mitigation of suicide risk so the data from this study showing that having a sense of achievement and involvement in school and parental oversight (knowing where a child is, what they are doing, and with whom) is a welcome addition to our knowledge.
From a methodological standpoint, the cross-sectional nature of the study means that we cannot make strong conclusions about the direction of the effects. However, the Adolescent Brain and Cognitive Development (ABCD) cohort study is a prospective study which means that, over time, the direction of these relationships may be uncovered.
A serious limitation of this study is the way the authors characterise and operationalise self-harmful behaviour. In this work (as in most research from the US) the authors treat self-harmful behaviour as a binary phenomenon. Either people who harm themselves have suicidal intent (therefore have attempted suicide) or they don’t (therefore they have engaged in Non-Suicidal Self-Injury, NSSI).
The reality, and extant data, suggests that the relationship is much more nuanced and that these behaviours are, in fact, strongly related. Indeed, taxometric studies reveal that self-harmful behaviour is dimensional in nature, not dichotomous (Orlando et al, 2015). Moreover, the intent associated with self-harmful behaviour changes over time in young people with intent emerging over time and with repeated episodes (Townsend et al, 2016).
Recent findings from a UK prospective cohort study has shown that NSSI significantly predicts the transition from suicidal thoughts to suicidal behaviours (Mars et al., 2019). Several studies now indicate NSSI as one of the most powerful predictors of future suicide attempts, positing NSSI as a ‘gateway’ to future suicide (e.g. Whitlock et al, 2013). Indeed, others have noted that several robust studies have now demonstrated that ‘… self-injury that is not suicidal in intent at the time is associated with future risk of suicide attempts and possibly suicide.’ (Wilkinson, 2011, p.742). Thus, it is vital that future interrogations of ABCD data include all facets of self-harmful behaviour; those with suicide intent and those without.
The study employed regression models to predict ‘suicidality’ and for child-reported suicidality only 24% of variance is accounted for in these models. This leaves the majority of the variance in childhood suicidal thoughts, plans and behaviours unaccounted for. In future prospective studies, it may be possible to include an even more comprehensive range of biopsychosocial factors from the ABCD data set to increase the variance explained. Including all facets of self-harmful behaviour (e.g. NSSI as a predictor of attempts) would also likely significantly increase predictive utility and variance explained (Mars et al, 2019; Whitlock et al, 2013).
The finding relating to weekend screen time use is interesting, but requires significant future investigation in order to understand the direction of the effects and the mechanisms that underpin them (e.g. cyberbullying vs. social withdrawal). Prospective studies are urgently required on this issue as there are likely to be both benefits and risks associated with engagement in the digital world (Marchant et al 2017).
Implications for practice
As others have noted (O’Connor et al, 2020), given the low concordance between child and parental reports of suicidality here, it is important that others closely involved in the lives of children are trained and skilled in identifying young people who are struggling and who may require support.
Given the importance of childhood psychopathology it is vital that suitable age-appropriate interventions are offered to those struggling with mental health problems. Universal school based and parenting interventions may be helpful in this age group (e.g. Youth Awareness of Mental Health programmes) (Wasserman et al, 2015).
Statement of interests
Janiri D, Doucet GE, Pompili M et al. (2020). Risk and protective factors for childhood suicidality: a US population-based study. Lancet Psychiatry, Vol. 7, No. 4, p292–293
Hawton K, Bale L, Brand F et al (2020). Mortality in children and adolescents following presentation to hospital after non-fatal self-harm in the Mulitcentre Study of Self-harm: A prospective observational cohort study. The Lancet Child and Adolescent Health 4(2) 111-120.
Orlando CM, Broman-Fulks JJ, Whitlock JL, Nonsuicidal Self-Injury and Suicidal Self-Injury: A Taxometric Investigation. (2015). Behav Ther. Nov;46(6):824-33. doi: 10.1016/j.beth.2015.01.002.
Marchant, A., Hawton, K., Stewart, A., et al (2017). A systematic review of the relationship between internet use, self-harm and suicidal behaviour in young people: The good, the bad and the unknown. PLOS ONE 12 8 e0181722
Mars B Heron J Klonsky ED et al. (2019) Predictors of future suicide attempt among adolescents with suicidal thoughts or non-suicidal self-harm: a population-based birth cohort study. Lancet Psychiatry. http://dx.doi.org/10.1016/S2215-0366(19)30030-6
NCISH (2017). Suicide by children and young people in England. National Confidential Inquiry into Suicide and Homicide by People with Mental Illness (NCISH). University of Manchester, Manchester.
O’Connor R and Robb KA (2020). Identifying suicide risk factors in children is essential for developing effective prevention interventions. Lancet Psychiatry, Vol. 7, No. 4, p292–293.
ONS: Office for National Statistics (2017) Deaths registered in England and Wales (series DR): 2016. Accessed 7-5-20.
Townsend, E. (2019). Time to take self-harm seriously. (Commentary) Lancet Psychiatry 6(4) 279-280.
Townsend E, Wadman R, Sayal K, et al. (2016). Uncovering key patterns in self-harm in adolescents: Sequence analysis using the Card Sort Task for Self-harm (CaTS). Journal of Affective Disorders, 206, 161-168.
Wasserman D, Hoven CW, Wasserman C et al (2015). https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(14)61213-7/fulltextSchool-based suicide prevention programmes: the SYELE cluster-randomised, controlled trial. Lancet 18;385(9977):1536-44. doi: 10.1016/S0140-6736(14)61213-7
Whitlock, J., Muehlenkamp J, Eckenrode J et al (2013). Nonsuicidal self-injury as a gateway to suicide in young adults. Journal of Adolescent Health, 52(4), 486-492.
Wilkinson P (2011). Nonsuicidal Self-Injury: A clear marker for suicide risk. Journal of the American Academy of Child and Adolescent Psychiatry, 50(8), 741-743.
WHO Suicide data https://www.who.int/mental_health/prevention/suicide/suicideprevent/en/ Accessed 7-5-20