Find out more

Why does this matter?

Prevention of severe mental health difficulties could transform outcomes for younger people. This research will help services identify who is at the greatest risk.

What do we want to find out?

This study aims to examine whether social and educational risk factors recorded before the end of year 11 can predict severe mental health difficulties in the following years.

What we will do

We investigate school-specific and demographic risk factors for young people who have severe mental health outcomes. We aim to identify priority groups of young people who could benefit from early intervention. We focus on English state secondary school pupils and ask which educational and social factors are associated with:

• Psychiatric related attendance to A&E after year 11?

• Unplanned psychiatric related inpatient admissions after year 11?

Background and aims

Prevention of severe and long-term difficulties associated with the onset of symptoms of a mental health condition has yet to become commonplace in healthcare, despite some successes e.g. for psychosis and eating disorders. Initiatives are more likely to succeed with evidence-based implementation strategies which target priority groups that may be at the greatest risk of severe outcomes such as A&E presentation or unplanned inpatient admission. In light of this, our research aims to identify potential priority groups for early intervention initiatives. Specifically, we study which educational and social factors are associated with English state secondary school pupils’:

1. psychiatric related attendance to A&E after year 11, and

2. unplanned psychiatric related inpatient admissions after year 11.

Methods and analysis

We will employ a retrospective cohort design using subsets of ECHILD data. The first cohort, for research question 1, will be people recorded in the National Pupil Database in the final term of year 11 between 2020 and 2021. Our index date will be end of year 11 (August 31st) and outcomes will be defined following the index date up to the latest available data using the Emergency Care Data Set (ECDS). The second cohort, for research question 2, will include people recorded in the NPD in the final term of year 11 between 2013 and 2022. Outcomes will be defined using the Emergency Care Dataset (ECDS) and Mental Health Services Data Set (MHSDS).

We will describe and summarise the population characteristics of this national cohort. We will fit statistical models appropriate to the research questions and the outcomes of interest (psychiatric related attendance to A&E after year 11 and unplanned psychiatric related inpatient admissions after year 11). Where the outcomes are count data, Poisson regression models or, in the case of continuous outcomes with overdispersion and zero-inflated counts (common properties of patient-level A&E presentation data), we will fit negative binomial regression models. Cox’s Proportional Hazards models will be fitted for binary outcomes. We will explore the effects of multiple exposure and confounder variables by fitting multivariate models. Statistical corrections for multiple testing will be used, alongside sensitivity analyses.

ECHILD data used

• HES: MHSDS, ECDS (2019-2023)

• NPD: Spring Census, Absences, Exclusions, Children in Need Census, Key Stages 2 and 4 (2014-2021)

Stakeholder and lived experience involvement

People with related lived experience of both childhood disadvantage and mental health problems were involved in the design, implementation, analysis, interpretation, and writing of the landing page text as part of a project working group. Experts in linked data analysis, children and young people’s mental health, and experts in policy research were included in the Project Working Group. The lead researcher has relevant lived experience.