The effect of low childhood income on self-harm in young adulthood: Mediation by adolescent mental health, behavioural factors and school performance

Abstract

Low childhood income is an established risk factor of self-harm in adolescence and young adulthood, and childhood income is additionally associated with various correlates of self-harm. How these correlates, such as psychiatric disorders, substance abuse, violent behaviour and school problems, mediate the effect of childhood income on self-harm, is less understood. The purpose of the current paper is to examine this mediation. The study is based on administrative register data on all Finnish children born in 1990–1995. An analytical sample of 384,121 children is followed from age 8 to 22. We apply the parametric g-formula to study the effect of childhood income on the risk of self-harm in young adulthood. Adolescent psychiatric disorders, substance abuse, prior self-harm, violent criminality and victimization, out-of-home placements, not being in education, employment or training and school performance are considered as potential mediators. We control for confounding factors related to childhood family characteristics. As a hypothetical intervention, we moved those in the lowest childhood income quintile to the second-lowest quintile, which resulted in a 7% reduction in hospital-presenting self-harm in young adulthood among those targeted by the intervention (2% reduction in the total population). 67% of the effect was mediated through the chosen mediators. The results indicate that increases in childhood material resources could protect from self-harm in young adulthood. Moreover, the large proportion of mediation suggests that targeted interventions for high-risk adolescents may be beneficial. To our knowledge, this is the first paper to use the parametric g-formula to study youth self-harm. Future applications are encouraged as the method offers several further opportunities for analysing the complex life course pathways to self-harm.

Neighbourhood‐related socioeconomic perinatal health inequalities: An illustration of the mediational g‐formula and considerations for the big data context

“…10 This technique is recognised as a unified and flexible causal inference approach that allows for designing custom interventions, a property available in a few other methods. 11,12 The g-formula was originally proposed for applications in settings with confounders affected by previous exposure and can naturally be extended to mediation analysis. 12 Mediation analysis evaluates the relative magnitude of pathways by which an exposure influences an outcome.…”

“…[16][17][18][19] The parametric g-formula extends mediation analysis to settings involving non-linearities and interactions, 10,12 and its estimates are easily understandable population-averaged effects. 11 The g-formula uses parametric regression models to predict outcomes under hypothetical intervention scenarios (counterfactuals), which are used to estimate mediation effects via Monte Carlo simulation. The parametric g-formula is referred to as mediational g-formula when used application can contribute to providing valuable insights for the development of policy and public health interventions.…”

“…However, its widespread application only became feasible with increasing computational power 10 . This technique is recognised as a unified and flexible causal inference approach that allows for designing custom interventions, a property available in a few other methods 11,12 . The g‐formula was originally proposed for applications in settings with confounders affected by previous exposure and can naturally be extended to mediation analysis 12 …”

“…The traditional mediation approach has important shortcomings as it is prone to bias when exposure and mediator interact and when the outcome is non‐linear (e.g., dichotomous) 16‐19 . The parametric g‐formula extends mediation analysis to settings involving non‐linearities and interactions, 10,12 and its estimates are easily understandable population‐averaged effects 11 . The g‐formula uses parametric regression models to predict outcomes under hypothetical intervention scenarios (counterfactuals), which are used to estimate mediation effects via Monte Carlo simulation.…”

Multilevel interactions between family and neighbourhood socioeconomic indices in childhood and later risks of self-harm and violent criminality in Denmark: a national cohort study

“… 1 Experiencing familial deprivation can negatively affect a child's socio-emotional, behavioural, and cognitive development, 2 , 3 and increase the likelihood of aggressive behaviour. 4 For example, some parents with a low income might be more likely to have psychiatric disorders, abuse substances, and have impaired parenting behaviours because of poverty-related stress. 5 The higher frequency and severity of such problems in these families increases the risk of both inward-directed (self-harm) 6 and outward-directed (violence) 7 aggression.…”