Recruitment of Families to ASD Research Databases through the UK Publicly Funded Healthcare System

Background:  

Large numbers of representative families of children with ASD are needed to make significant advances in some research areas. However, researchers frequently struggle to recruit adequate numbers of participants, and some families have more opportunity to contribute to research than others.

Objectives:  

1. Present the database recruitment methods (including the challenges identified)
2. Provide evidence that those registered are representative of UK families of children with ASD
3. Demonstrate the utility of collecting data from all families as they join
4. Show how in subsequent projects, we can recruit very large number numbers of families from the databases, allowing effective subgroup analysis

Methods:  

The Autism Spectrum Database-UK (ASD-UK) and the Database of Children with ASD living in North East England (Daslⁿe) aim to increase researchers’ access to families, and families’ access to research studies. Daslⁿe has a population sampling frame and commenced recruitment from multidisciplinary publicly health and education services in 2003; it now includes 1450 families (around 55% of local families). ASD-UK started recruiting through the National Health Service in 2011 and has recruited 1350 families. Following approval by the databases’ research committee, almost 3000 families can be contacted by UK ASD researchers. Data about child and family characteristics are collected from parents and clinicians. Daslⁿe and ASD-UK have now supported more than 30 UK research projects.

Results:  

Representativeness: Characteristics of children recruited to ASD-UK and Daslⁿe are similar, and representative of families with ASD in the UK on gender, DSM-IV ASD diagnosis, schooling type, and proportion with learning disability. Children of parents who declined participation in ASD-UK and those whose parents consented are very similar in year of birth, ASD diagnosis, gender and social deprivation score. Thus database families are representative of ASD families, and not a biased sample.

Utility of collecting data when parents join:

1. Analysis of over 2100 families showed that children’s median age at ASD diagnosis has not reduced over recent years, and fewer than one fifth of children were diagnosed before age 3 years (Brett et al., in preparation).

2. An investigation of recurrence risk for ASD revealed that 7% of 2300 families have two or more children with ASD. Subgroup analysis revealed this rose to 10% when only families who have finished having children were included. There was strong evidence for an association between reproductive stoppage and ASD (Wood et al., 2014).

Effective subgroup analysis can be undertaken through large research databases: Koshy recently recruited 650 families to the ASD+ study of co-existing conditions in children using postal questionnaires. More than two-thirds of families had at least one clinical unmet need. Parents of children with moderate to severe impairment in co-existing emotional and behavioural conditions had greater unmet needs than parents of children with no/minimal impairment.

Conclusions:  

Through publicly funded UK health and education teams, very large numbers of children and families representative of the overall ASD population can be recruited. Large samples are needed to collect high quality data and undertake subgroup analyses in relation to children’s ASD/family characteristics.