ADR UK Research Fellows: Education and Child Health Outcomes from Linked Data

This project aims to explore the following research questions:

1. How many children present at hospital with an injury or illness indicative of acquired brain injury, and how does the prevalence of acquired brain injury vairy across different socio-demographic profiles?  
2. What special educational needs provision do children with acquired brain injury receive in schools?   
3. Are children with acquired brain injury more vulnerable to school exclusion, persistent absence, mental health difficulties, or poorer educational attainment?  
4. How does special educational needs support moderate the impact of acquired brain injury on these outcomes?  

The methodology used in this study:  

This study will analyse linked health and education data for all children born between September 2003 and August 2004 who attended a state-funded school in England. Hospital data will be used to identify any illness or injury indicative of an acquired brain injury (for example traumatic injuries, encephalitis, stroke, brain tumours). The research will try to understand the severity of these injuries from available hospital records, and will then look at education data to track outcomes longitudinally.  

The methods used in this study will include:  

• Tabular and graphical descriptive statistics, to understand the sample, and simple bivariate statistical tests to see whether there are differences in acquired brain injury prevalence between sociodemographic groups.   

• Regression models to assess how outcomes (like being identified for special educational needs support) are impacted by the presence of acquired brain injury, and to assess whether sociodemographic profiles or geography are impacting these outcomes.

• We will complement the regression analysis with treatment effects methods for observational data (showing the difference in outcomes between a treatment group and a control group). This will ensure our analysis is robust against possible biases. We will use matching estimators including ‘Nearest Neighbour’ matching, and Propensity Score Matching. 

Funded value:  £194,097 

Duration: October 2024 – April 2026 

This project aims to explore the following research questions:

1. Are there regional or geographical disparities in adverse outcomes among pupils with neurodivergent special educational needs? 
2. What risk factors are associated with medically-related absenteeism, self-harm, and suicide among pupils with neurodivergent special educational needs?  
3. What risk factors, especially health care utilisation, are associated with alternative provision and formal school exclusion among pupils with neurodivergent special educational needs? 
4. Is it possible to identify the causal relationship between special educational need provision and suicide or self-harm among pupils with autism spectrum disorder? 

The methods used in this study will include:

• Descriptive statistics will be produced to understand rates of adverse outcomes over time among pupils with neurodivergent special educational needs. Spatial analysis will also be used to understand any geographic variations in these outcomes. Because the data in this project is clustered (e.g. pupils attending the same school, schools based in the same local authority, etc), analyses will use multilevel modelling to account for this built-in structure.
• For binary outcomes (e.g. whether a pupil was formally excluded or not), logistic regression will be used to assess associations with risk factors; for outcomes which are counts (e.g. number of days absent due to a medical reason), Poisson or negative binomial regression will be used  

• Data can often be missing or incomplete; this project will consider ways of dealing with missing data including complete case analysis (i.e. only analysing individuals for which all complete data is available) and multiple imputation, a technique for imputing or guessing at missing data.  

• Attempting to identify a causal relationship between differential special educational needs provision and suicide or self-harm among pupils with autism spectrum disorder will use an approach called target trial emulation, a type of study design which seeks to simulate a randomised clinical trial using observational data.  

Funded value: £199,660.51 

Duration: October 2024 – March 2026 

This project aims to explore the following research questions:

1. What proportion of young people experiencing repeated absence, poor attainment, or school exclusion have underlying chronic mental or physical health conditions documented in hospital records? 
2. How much of the association between family disadvantage and educational outcomes is attributable to family risk factors or underlying chronic mental or physical health conditions in young people that can be identified within the ECHILD dataset?  
3. Has the impact of health on educational inequalities changed since the pandemic?  

The methods used in this study will include:

• To identify the characteristics of young people at risk of poorer educational outcomes, this project will calculate the proportion of persistent absence, low educational attainment, and exclusion within the ECHILD cohort. This will be done by deprivation level, history of maternal mental health difficulties, whether there has been social service contact with the family and whether the young person has a chronic mental and/or physical health condition (and its severity).
• The project will examine how much of the association between family disadvantage and educational outcomes is attributable to family risk factors or underlying chronic mental or physical health conditions in young people. It will then examine data over time to identify whether the impact of health on educational inequalities has changed since the pandemic. 

Funded value: £139,897 

Duration: May 2025 - November 2026 

This project aims to explore the following research questions:

1. What factors influence whether children referred to Child and Adolescent Mental Health Services (CAMHS) receive a diagnosis of ASD? 
2. How does the likelihood of receiving a diagnosis differ based on factors like socioeconomic background and ethnicity? 
3. How does the support provided for special educational needs in schools affect the relationship between receiving an ASD diagnosis in CAMHS and children's health and education outcomes? 
4. How do these effects above vary depending on factors like socioeconomic background and ethnicity? 
5. How can the services provided by CAMHS and special educational needs support be improved based on the findings of this research, to help reduce differences in outcomes for children from different backgrounds? 

The methods used in this study will include:  

• This project consists of three connected work packages, using a wide range of research methods including advanced statistical analysis and actively engaging multiple stakeholders. 
• The first work package aims to find de-identified data on groups of children diagnosed with ASD from referrals in CAMHS, and understanding what factors predict this diagnosis, including socioeconomic and ethnic differences. Eligible children will be followed from 2016 to 2022 to reveal various factors related to their diagnosis, such as why they were referred, the type of service they received, and their age at diagnosis. This will help us understand how many children are diagnosed with ASD, their characteristics, and any inequalities. 
• The second work package intends to understand how being diagnosed with ASD and receiving special educational needs support at school affects children's health and education outcomes. To begin with, this work package will find out how being diagnosed with ASD affects the level of support children receive at school, and how this varies by ethnicity and socioeconomic background. Additional research can then be conducted to find out how ASD diagnosis and special educational needs support affect children's exam scores and use of mental health services. 
• The third work package explores why some children with ASD have worse health and education outcomes than others, and how to reduce these inequalities. This can be achieved by working with local health and education services and ASD communities to understand how these systems work and how decisions are made. By identifying gaps in the current systems, suggestions can be made on ways to improve support for children with ASD and reduce inequalities in their outcomes. Such findings will be shared with local services and communities to help them make informed decisions about supporting children with ASD. 

Funded value: £174,852

Duration: September 2024 – June 2026

This project aims to explore the following research questions:

1. What factors drive mothers’/birthing parents' engagement with mental health services before and during pregnancy? 
2. What factors influence mothers’/birthing parents' use of emergency department services during pregnancy? 
3. Are perinatal mental health issues and emergency department visits during pregnancy linked to subsequent child health, development, and educational outcomes? 

The methods used in this study will include:

First, some basic facts about the mothers, parents, and their children will be gathered using the de-identified data, such as ages or other characteristics like the region they live in. Simple mathematical tools (like averages and percentages) will be used and charts will be created to understand this information more clearly. It will also be possible to analyse geographical differences to see if certain areas have different outcomes in terms of health or education. 

Next, more detailed analyses will be performed to see if we can predict certain outcomes. For example, it will be possible to check if mothers who visited the emergency room during pregnancy are more likely to have children who do well in terms of their development and early education. Methods will also be used to break down the data into groups who might be more or less at risk and find patterns that could help us predict these outcomes more accurately. 

Finally, these prediction models will be assessed to see if they can be used in real life to identify mothers or children who might need extra support. The project may simulate a clinical trial to better understand how accessing mental health services affects the health and wellbeing of both mothers and their children’s development and educational outcomes. 

Funded value: £188,061 

Duration: January 2025 - July 2026 

This project aims to explore the following research questions:

1. When do neurodivergent children and young people who develop a mental health condition first engage with any existing educational support, social care, and health services?
2. When do these children first engage with each type of educational support, social care, and health services?
3. Does the time to and likelihood of service engagement (overall and for specific services) differ between neurodivergent and neurotypical children and young people who develop a mental health condition
4. Does the time to and likelihood of service engagement (overall and for specific services) differ between neurodivergent children and young people who develop a mental health condition and those without a mental health diagnosis?

The methods used in this study will include:

This project research project will be delivered in three stages:

1. The first stage will involve data exploration and recoding. The quality of the information recorded in the Mental Health Services Data Set will be assessed to determine if it could be used to define each child’s neurodivergence and mental health diagnosis status. If not, social and education data will be used to complement this information. Three comparison groups will be identified:
   1. neurodivergent with a mental health diagnosis
   2. neurodivergent without a mental health diagnosis
   3. not neurodivergent but with mental health diagnosis.
2. Descriptive statistics will then provide a first insight into the data and preliminary comparisons across groups. This analysis will provide a general overview of differences in the type of services children and young people are in contact with and how that varies based on demographic characteristics (gender and ethnicity).
3. Time ­to ­event analysis (or survival analysis) is a collection of statistical procedures that estimates the amount of time it takes before a particular event of interest occurs. The project will use this analysis to estimate and compare how long it takes for the children and young people in different groups to first come into contact with any support services and then how long it takes for each support type.

The project will actively engage lived experience advisors (such as neurodivergent young people and their parents) and policymakers (such as local commissioners) throughout the research lifecycle.

Funded value: £199,982

Duration: October 2024 – March 2026