PhD researcher publishes in Social Psychiatry and Psychiatric Epidemiology

PhD researcher publishes in Social Psychiatry and Psychiatric Epidemiology

PhD researcher publishes in Social Psychiatry and Psychiatric Epidemiology


PhD researcher Amanda Spikol and her supervisors, Professor Jamie Murphy and Dr Donal McAteer, have published the results of a study on autistic trait behaviour in Social Psychiatry and Psychiatric Epidemiology; ‘Recognising autism: a latent transition analysis of parental reports of child autistic spectrum disorder ‘red flag’ traits before and after age 3’. This study was the basis for her undergraduate dissertation.

Often parents (especially new parents) find their worries trivialised or brushed aside by GPs and many in contemporary media eschew the idea of early autism spectrum disorder (ASD) screening. However, while a wait-and-see attitude can seem prudent, early intervention programs have a clinically-proven track record of difficulty reduction; potentially changing the life of a child on the spectrum. The question remained: Are parents’ observations of ‘red flag’ behaviours before age 3 indicative of disorder and are they predictive autistic trait behaviour severity in later childhood? This study utilised a large population data sample, part of a multiphasic census conducted in by the Office for National Statistics (ONS) in 1999/2004, to examine the accuracy of parental concerns over their children’s behaviour.

Taking from the 2004 cohort (N=7977), the team used 5 anchor questions from an autism survey designed by the ONS, which asked parents about behaviour problems in 5 key domains <age 3. Then, they isolated 10 questions from the body of the survey concerning the same domains >age 3. A latent class analysis (LCA) was conducted to determine behaviour-typified profiles within the <age 3 sample, resulting in Low, Moderate, and High endorsement classes. A second LCA was conducted for the >age 3 data, also resulting in Low, Moderate, and High presentation classes. The team chose several variables associated with ASD (epilepsy, learning disability, poor coordination, anxiety, and an ASD diagnosis) as external validators of the latent factor in a multinomial logistic regression. Put plainly, it was hypothesised that the differences in child behaviour were due to an unseen-but-assumed factor and that factor was ASD. Compared to the Low group, the Moderate and High group showed significantly high odds of also having one or more of the variables known to be associated with ASD.

Following, a quasi-latent transitional analysis (qLTA) was run between the <age 3 and >age 3 data in a move-or-stay model to determine if children were transitioning to different classes over time, or remaining in place. A majority of the overall sample (89.9%) ‘stayed’, meaning that Low endorsement children <age 3 became Low presentation children >age 3, and so on. Of the ‘movers’ in the model, there was a downward trend as 72.7% transitioned from High to Moderate or Moderate to Low. This was in keeping with the ‘optimal outcome’ model. These results highlight that parents who identified ‘red flag’ traits <age 3 were validated in their concerns, as these behavioural traits remained stable as years passed. Also of note was that most transition was ‘down’ in severity, showing that traits which are at first problematic can improve. In this population, parental perceptions of these behaviours were reliable indicators of ASD risk in later childhood. The implications of these findings reinforce that parents are often the most reliable observers of their children’s behaviour and can be the first to raise concerns and initiate early intervention programs. The paper, which features tables and lovely colourful graphs, is available here.