Behavioural Development, Long-term Outcomes and Opportunities to Optimize Youth Mental Health Trajectories

We propose an accelerated longitudinal study design that allows covering a long developmental period (9 to 25 years) in a short study time (5 years follow-up). Youth and family caregivers will be recruited from various sources and followed-up annually for the first five years. Beyond this 5-year period, further follow-ups will depend on continued study funding and may be either annual or less frequent (e.g. every 3-4 years), and either interview-based or restricted to administrative data. The present study includes assessment of antecedents, opportunities and outcomes that will establish eligibility for preventive interventions; any randomised intervention studies based on this cohort will be based on the outcome for specific REB submissions in the future.

Predicting the onset of DMDs using a set of readily accessible antecedents

We hypothesise that youth who develop DMDs will have higher severity and frequency of antecedents compared to those who do not develop this outcome. We anticipate that a clinically meaningful prediction (i.e., with accuracy 80% or more said to be good or excellent, while less than 60% is deemed not meaningful; many existing predictors perform at 60-80% range9) of DMDs can be made with the set of readily measurable antecedents. We will estimate the strength of antecedent-outcome relationship separately in the two (help-seeking or non help-seeking) subgroups. While we expect variations in the rates of outcomes (help-seeking > non-help seeking), we hypothesise that the relationship between antecedents and DMDs will be similar across the groups with no major drop in the ability to predict outcome in any subgroup (i.e. >60% accuracy in each, below which prediction is unlikely to be of any clinical benefit59).

Relationship between predictive antecedents, resilience and biobehavioural markers

In the consenting subsample, we will assess if the polygenic risk of individual disorders/pathways and summary measures of brain health (age-appropriate structure, chemisty, function) differ between those who develop DMDs and those who do not. We will subsequently test if measures with highest effect size differences, if any, will improve the accuracy (discrimination index) for outcome prediction by at least 10% to assess their suitability for routine clinical implementation.