Emotional and behavioural dysregulation markers in late childhood and early adolescence as risk factors for mental illness and healthcare utilisation in young people at the age of 18-24

Date of Approval
Application Number
Technical Summary

The incidence of mental illness in children and young people is reportedly rising, with one in eight having a diagnosable disorder. Furthermore, 75% of mental illness starts before the mid-20s and 50% before the mid-teens. Younger age at psychopathology symptom onset is a significant predictor of functional impairment in later life. Adolescence is therefore a crucial time for early recognition and intervention. Emotional and behavioural dysregulation has been shown to underlie several types of psychopathology in children and adolescents. However, adolescents’ emotional states also show considerable normal variation. Distinguishing early signs of, for example, depression or anxiety, from normal mood variation at this age can be challenging. However, few studies have examined how these symptoms evolve over the course of adolescence to become mental disorder. Studies are needed to assess the likelihood that help seeking in Primary Care for emotional and behavioural dysregulation in adolescence are early signs of mental illness requiring intervention.

We will use a retrospective cohort design using CPRD data to investigate if primary care consultations for mental health symptoms in late childhood and early adolescence (8-14 years) predict mental illness and healthcare utilisation in young adults aged 18-24 years.
Using logistic regression, we will examine whether emotional and behavioural dysregulation markers (such as anxiety, emotional upset or behavioural problems) in late childhood and early adolescence predict later mental health diagnoses, use of psychotropic medications, mental health referrals, hospitalisations and emergency department attendance for mental health reasons (e.g. for attempts of suicide such as overdoses) in young adulthood.
Using Poisson regression, we will address if these early consultations predict the frequency of healthcare utilisation.

Health Outcomes to be Measured

Our Primary outcome is likely mental illness at age 18-24 years. We will define likely mental illness as one or more of the following:

1) Specified diagnosis mental illness Read code (as defined in Appendix B); number of episodes (e.g. depressive episodes with evidence of remission between). Age group will be employed to identify age specific diagnoses

2) Use of psychotropic medications (e.g. antidepressants, anxiolytics, antipsychotics) for ≥6 months (first prescription to first discontinuation for each patient). We will also analyse, if available, time to treatment initiation (age at diagnosis to age at first prescription).

3) Referral to specialist Mental Health services: reason for referral, type of healthcare setting (e.g. IAPT, outpatient mental health, inpatient mental health)

4) Attendance at A&E for mental health reasons (e.g. psychotic episode, attempted suicide, non-suicidal self-injury).

Health resource utilization:
We will use records that measure healthcare utilisation (e.g. frequent attenders, referrals to mental health settings, emergency visits…). We will take into account:
1) Frequency of contact with primary care for mental health or physical symptoms
2) Frequency of referrals to mental health services
3) Frequency of mental health crisis presentations e.g. A&E visits for overdoses or non-suicidal self-injury (NSSI)), Mental Health Act assessments


Dasha Nicholls - Chief Investigator - Imperial College London
Ana Pascual Sanchez - Corresponding Applicant - Imperial College London
Dougal Hargreaves - Collaborator - Imperial College London
Geva Greenfield - Collaborator - Imperial College London
Hanna Creese - Collaborator - Imperial College London
Sonia Saxena - Collaborator - Imperial College London


HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation