The common childhood disorders that have been left out in the cold

Taxi drivers have become the modern equivalent of the man on the Clapham omnibus. I've conducted a totally unsystematic but long-term survey of them and can report that most know what autism and dyslexia are, but very few have any idea about specific language impairment (SLI). Now this is odd because SLI is reckoned to be as common as dyslexia and 10 times as common as autism.

In SLI, children have noticeable problems with understanding and/or producing language, for no obvious reason. However, it is clearly a condition with an image problem. So we have what I call the taxi driver paradox: how come taxi drivers (and lay people in general) know about dyslexia and autism but not about SLI?

The same paradox is seen in research activity. Three years ago, I attended three UK-based conferences that focused on autism, dyslexia and SLI respectively. The autism meeting was huge, the dyslexia meeting large, and the SLI meeting much smaller. Somehow, SLI just keeps getting overlooked. When I delved into possible explanations, I found some logical reasons for this, but they did not fully account for the neglect. Other more arbitrary factors, especially professional prestige, seem to play a worryingly large role in determining which disorders get attention.

When I first noticed the discrepancy, I decided to quantify my impressions by doing a bit of Googling, and so was born an intriguing project that led to a paper published today in the journal PLoS One. I decided that to gain real insights into factors affecting research activity, I should look more broadly at neurodevelopmental disorders, including some rare genetic disorders. I found a precedent for such a study in the BMJ, in a paper looking at neurological diseases. It concluded that, if research was quantified as a "publication ratio" reflecting the number of research publications per number of affected individuals, then common disorders, such as migraine, were under-researched relative to rare conditions such as Creutzfeldt-Jakob disease.

My analysis of neurodevelopmental disorders found exactly the same pattern. The average "publication ratio" was much higher for rare than for common disorders. But there was an obvious reason for this. The rare disorders were typically more severe and so I needed to take that into account. Clinical colleagues came to my aid by rating neurodevelopmental disorders on an ad-hoc four-point severity scale. Disorders also varied in the extent to which research involved genetics. Using a keyword search, I got a rough percentage of genetic research.

Taking into account the frequency of the disorder, its severity, and the proportion of genetic studies, my prediction of the amount of research devoted to a condition was pretty good, accounting for 73% of the variance. Nevertheless, there were still anomalies. The amount of autism research was pretty much in line with the prediction, given its frequency and severity, but both SLI and dyslexia were under-researched. Even more strikingly neglected were the conditions of developmental dyscalculia, developmental coordination disorder, and speech sound disorder. In contrast, the amount of research on ADHD and Tourette syndrome was high in relation to their frequency and severity.

An analysis of the change in the amount of research over time revealed further differences. There were 1,215 new publications on autism during 1985-1989, rising to 7,158 in 2005-2009. The increase for ADHD was even more dramatic, from 356 publications in 1985-1989 to 6,158 in 2005-2009. These rates of increase were startlingly higher than any of the other disorders under consideration.

How far were these differences in research rates being driven by public funding? To look at this I used a web-based resource operated by the main US funding body, the National Institutes of Health. This confirmed a very strong relationship between the amount of NIH funding for a disorder and the amount of research on that disorder. So the growth in research on autism over the years was paralleled by a staggering increase in funding, including some specific initiatives. The Combating Autism Act passed in 2006 by the United States Congress, authorised nearly $1bn (£640m) over five years to combat autism and related disorders.

One factor behind this is the increase in autism diagnosis over the past 20 years, which has created a sense of urgency. The increase is, however, at least in part due to changing diagnostic criteria: some individuals with the diagnosis have such mild symptoms that they challenge the notion that they have a disorder. Another factor is lobbying by special interest groups. These have been remarkably effective in getting autism on the agenda of policiticians. While this is admirable, it begs the question of why other disorders are left out in the cold.

It's not that autism and ADHD are somehow more tractable for researchers than other disorders. The issues being investigated are pretty similar, the different disorders raise the same questions about causes and, indeed, there are no sharp boundaries between the diagnoses. Most affected children have symptoms of several conditions and few meet the criteria for just one diagnosis.

A final analysis in my PLoS One paper suggests yet another reason for the differential interest in different disorders, and a worryingly unscientific one: the status of the profession most closely associated with the condition. The NIH funding analysis showed that disorders differed not just in the amount of funding they attracted, but also in the source of that funding. ADHD, autism and Tourette syndrome were categorised by NIH under mental health, dyslexia was put under child health and development, and SLI under communication disorders. The professional disciplines associated with each category differ substantially in prestige: medicine for mental health, psychology for child development, and speech and language therapy for SLI.

My concern is that an essentially arbitrary compartmentalisation of conditions in relation to professional disciplines appears to have led to a gross inequality in how closely related conditions are perceived and investigated.

Dorothy Bishop is a professor in developmental neuropsychology at the University of Oxford and blogs at BishopBlog