Core research question:

Does identification/diagnosis of children with attention deficit hyperactivity disorders in their communities help improve core symptoms?

Background to proposed research:

Biomedical research suggests that attention deficit hyperactivity disorder (ADHD) is partially genetically and neurologically determined, leading to recommendations that a clinical diagnosis should be made as early as possible to treat and manage this condition. Current clinical trends show diagnosis is given more frequently and at younger ages than previously. Although the evidence for the effectiveness of the pharmaceutical intervention methylphenidate is well established as an effective intervention for ADHD, its use is controversial, particularly in young children.

There is little evidence-base that early diagnosis confers an advantage. Specific early school-based screening and intervention programmes for ADHD have not consistently improved children's outcomes.My own previous work in collaboration with Jean Golding and Colin Steer has shown that children with autistic traits in the ALSPAC cohort who were diagnosed had worse outcomes as adolescents in core autistic symptoms than those who had equally severe symptoms as preschoolers, but remained undiagnosed. We were not able to follow the developmental trajectories of core autism symptoms in detail other than in measures prosocial behaviour, where diagnosis and subsequent interventions in the community apparently made no difference to the developmental trajectory of this trait.

The proposed research will extend the methodology used for autistic traits to examine ADHD. We will identify ADHD diagnoses from clinical medical records of children with SEN or parent and teachers reports of doctors diagnoses of ADHD, and the group of children in the cohort reported as having an ADHD diagnosis will be identified.

This will allow us to define a control group with similar ADHD symptoms levels to those with a diagnosis. - as recorded by parents, teachers and/or medical records- see above. We will then examine social and demographic factors that may be associated with lack of diagnosis, as well as comparing outcomes between the groups.

Methodology:

The main study will focus on four key areas, although there will be scope for flexibility to allow for new developments:

[A] Identifying behavioural traits most predictive of receiving a diagnosis.

Hypothesis:

There will be a group of impaired (but undiagnosed) children in the ALSPAC cohort.

This will involve backwards 2 stage logistic regression to find behavioural traits associated with receiving a clinical diagnosis, then defining a composite 'attention deficit, hyperactivity' trait with appropriate weightings, which matches diagnostic criteria as closely as possible. We will find the ADHD diagnoses from a combination of parent report, teacher report and the IDI data which reports ADHD diagnoses for children with SEN at school action plus or statement level. We may also utilise the DAWBA 'pseudo diagnoses' where children were rated as having ADHD, and determine what proportion were actually diagnosed with ADHD in their communities.

[B] Defining an undiagnosed control group with equivalent symptoms to those with diagnosis.

The group will consist of children with impairment at the same levels as those with diagnosis (in the top 5%, of a composite 'attention deficit, hyperactivity' trait, for example) but who have not been given any Special Educational Needs (SEN) provision. We will examine the sensitivity and specificity of the composite ADHD trait in predicting diagnosis, and we will identify an impaired (control) group who did not receive SEN provision.

[C] Looking at social and demographic factors to see if any predict diagnosis.

Hypothesis:

Social and demographic factors will influence access to diagnosis.

Regression/association will be used to determine which factors in a child's background predict whether a diagnosis is applied or whether they fall into the control group. Gender, geographic region, measures of socio economic status, maternal age at birth, maternal mental health and other factors will be examined.

[D] Comparing outcomes in the two groups up to adolescence, and with the general population of the cohort.

Hypothesis addressing core research question:

Diagnosis of ADHD and subsequent intervention has an effect on clinical and social outcomes, compared with the undiagnosed group with similar symptoms.

We would hope that children receiving diagnosis of ADHD and receiving evidence based treatment (stimulant drugs and / or SEN provision) will have better outcomes than children who were not diagnosed and did not have SEN provision. This stage will compare measures of core symptoms, in diagnosed group, or not diagnosed control group, and in the general population. In addition, we will look at psychological measures such as well being, bullying, self esteem, friendships, and other salient measures such as impact on carers and children as the children reach adolescence. Trajectories of traits throughout the children's lifetimes will also be compared where data resolution allows.

The research proposed for this fellowship will take place at the host institution (Peninsula College of Medicine & Dentistry, Exeter), with an initial placement of one year at the University of Bristol Centre for Child and Adolescent Health, focusing on epidemiological and statistical techniques with ongoing collaboration on the qualitative project with the school of Social Science at Exeter University. A second period of up to one year (broken into short chunks) will be a placement at the international centre for epidemiological research with the larger longitudinal cohort in Denmark, depending on agreements and data availability. Preliminary enquiries have been made to The Danish National Birth Cohort (n=100,000+) who seem enthusiastic to collaborate.

The proposal as a whole responds to calls for community-based, multidimensional studies of outcome in childhood disorders so that current clinical practice can be robustly married with evidence based research. Please note that this research will be part of a larger fellowship proposal made to the Wellcome trust. The deadline for submission of this proposal is 21st January, so I am hoping to secure agreement with ALSPAC before submitting it.