Aim 1: Identification of high-risk group

A core concern is to ensure that resilience is defined by reference to a clearly defined risk, and that variation in child outcomes does not merely reflect variation in severity of exposure to the risk factor of interest. Parent depression symptom data from pregnancy up to child age 12 will be analysed using group-based trajectory models to identify parents with chronic high levels of depressive symptoms. Additional information on severity of comorbid psychopathology, and functional impairment will be used to further characterise variation in severity of risk.

Aim 2: Definition of offspring mental health resilience

Most previous studies have focused on specific outcomes or on adaptation on a single occasion, limiting conclusions about the extent and stability of positive adaption in children defined as resilient. The second aim will be to define resilience by assessing adaptive youth outcomes over multiple mental health and related psychosocial domains (i.e. absence of depression, anxiety, disruptive behaviour disorders, substance use, self harm, suicidality, lack of functional impairment as rated on SDQ and DAWBA) and over multiple time points spanning adolescence. Sensitivity analyses will address how far results generalise across varying definitions of resilience.The project will also attempt to go beyond simple categorical indicators of resilience using latent growth mixture modelling in order to distinguish adaptive from maladaptive developmental pathways.

Aim 3: Identifying predictors of resilience in children of depressed parents

The third aim will be to evaluate the extent to which the different parent, family, child and peer factors outlined above predict resilience. Analyses will assess longitudinal associations using multivariate models to highlight the most important independent predictors of resilience in this high risk group.

Aim 4: Testing causal mechanism underlying resilient adaptation

Having identified the most important predictors of resilience, the study will use structural equation modelling to further exploit the longitudinal nature of the two studies and test causal mechanisms underlying well-being. For example, meditational models will assess whether resilience arises when key risk pathways are interrupted, and moderational models will test which factors act as buffers to enhance children's resilience. Cross-lagged models of change will assess direction of effects, and analyses will consider potential confounding variables.

Aim 5: Identification of predictors of resilience of particular relevance to children of depressed parents

The ALSPAC data offers the opportunity to distinguish between general predictors of positive mental health and specific predictors of resilience in children of depressed mothers. In particular, we will test interactions between resilience factors and maternal depression risk status.