Evidence weak for social communication disorder

There are several reasons why social communication disorder should not have been included in the DSM-5, says Helen Tager-Flusberg.

By Helen Tager-Flusberg
30 May 2013 | 6 min read
This article is more than five years old.
Neuroscience—and science in general—is constantly evolving, so older articles may contain information or theories that have been reevaluated since their original publication date.

As part of its overhaul of the classification of autism and related pervasive developmental disorders, the DSM-5, the latest edition of the Diagnostic and Statistical Manual of Mental Disorders, introduces a new category: social (pragmatic) communication disorder (SCD), classified under the heading of communication disorders.

The new disorder’s features include difficulties with social uses of both verbal and nonverbal communication, which overlap with the social communication impairments that now define autism. The major distinction between SCD and autism is that SCD does not include repetitive behaviors or restricted interests, thus capturing a group of children that would fall outside autism’s definition.

In the public eye, changes to autism’s diagnostic criteria have sparked the most attention and controversy, but among clinical researchers, SCD is raising far greater concern. The essential problem, as many commentators see it, is that there is little evidence that SCD has either reliability or validity as a distinct category of neurodevelopmental disorder1, 2, 3.

No one doubts that there are children and adults who have significant problems with social communication and pragmatic aspects of language, but the question is, do these problems warrant a separate clinical diagnosis or are they better viewed as comorbid symptoms seen in individuals with other neurodevelopmental disorders?

Like the framers of the DSM-5, I too once wanted to believe in the existence of SCD (or, as it is referred to in the literature, pragmatic language impairment or semantic pragmatic disorder). Indeed, my colleagues and I received a grant in 1998 to investigate this class of language disorder that seems to lie somewhere between autism and specific language impairment (SLI).

The trouble was that we never found any children who met the criteria for pragmatic language impairment who didn’t also have either autism or SLI, especially when we dug into their developmental histories.

Multiple reasons:

On reflection, then, there are several reasons why SCD should not have been introduced into the DSM-5.

First and foremost, SCD opens up a new Pandora’s box — just when the DSM-5 had finally dealt with categories that had been so troubling in the DSM-IV, namely Asperger syndrome and pervasive developmental disorder-not otherwise specified.

SCD was meant to fill a gap, perhaps, for children who don’t quite meet the new stringent criteria for diagnosing autism. But it will clearly not help clinicians who are faced with diagnosing a child who shows all the criteria for social communication impairments but only one symptom of restricted or repetitive behavior patterns.

Second, there doesn’t seem to be any evidence for the existence of a disorder that is genuinelyindependent of autism and SLI. Yes, there are children who, by the time they reach school, no longer show any restricted or repetitive behaviors and whose social communication impairments are mild enough that a clinician would say they no longer meet criteria for autism (though they clearly once did).

There are also children who may have been significantly delayed in acquiring vocabulary and grammar as toddlers and preschoolers (which would have warranted a diagnosis of SLI), and who still have problems with pragmatics, even though their vocabulary and grammatical skills are now within the normative range.

But the histories of such children reveal that they would have once met criteria for these other disorders. From the perspective of how to treat these children, it would be better if they retained their historical diagnoses so that they would be eligible to receive the evidenced-based interventions that address the social, communication and residual language problems that they still have.

We still lack reasonable clinical measures for assessing pragmatic language impairments. There are a few standardized tests on the market, but many children who clearly have quite serious problems communicating in everyday life can score well on these tests, as they tap meta-pragmatic knowledge (knowing the right answer in a test situation), not on-line communication skills (the ability to use that knowledge in a real-life situation).

Instead, researchers and clinicians rely on parent or teacher report measures (for example, the excellent Children’s Communication Checklist-2, developed by Dorothy Bishop) or the laborious, detailed analyses of transcribed conversational speech for which there are no normative data available for comparison. Most clinicians are not trained in identifying pragmatic impairments, so the absence of readily accessible assessment instruments means that they don’t have the tools with which to make a diagnosis of SCD.

Cultural bias:

All clinical diagnoses are subject to cultural bias or interpretation, a point that is emphasized in the DSM-5. But SCD is probably even more vulnerable to this kind of bias than any other disorder. Sociolinguists have documented how differently people communicate in different cultures.

Even in our own society, contrasting patterns of communication (‘Mars and Venus’) are among the most replicable of all sex differences. The examples of problems listed in the definition of SCD could so easily be seen in immigrants or English-as-a-second-language learners (mis)leading an insensitive clinician to label a child with a clinical disorder who simply comes from a different cultural background.

My final concern is that the age criterion for diagnosing SCD is vague and even contradictory as currently written in the DSM-5. On the one hand it states that “onset is in the early developmental period” — but what is early? Later we read “diagnosis … is rare among children younger than 4 years” and then “milder forms … may not become apparent until early adolescence.”

I suspect that what underlies these statements is the fact that there is flimsy research literature on SCD, with just a few studies available, none of which included children under age 4. I might take this one step further to suggest that the vagueness reflects the fact that SCD may not really exist as a distinct and unique disorder at any age.

Still, deficits in social communication areimportant for clinicians to identify. There is a growing literature to suggest that they are a significant problem for individuals who outgrow (usually as a result of good treatment) their earlier diagnoses of autism or SLI, for children with other learning disorders or attention deficit hyperactivity disorder, as part of the phenotype for genetic disorders such as Williams syndrome or fragile X syndrome, and even as a core symptom in schizophrenia.

I would argue that social (pragmatic) communication impairments are symptoms that should be evaluated as comorbidfeatures across a wide range of children and adults with other disorders. However, it was, in hindsight, a mistake for the DSM-5 to include SCD as a new category of neurodevelopmental disorders without carefully considering whether it has any reliability or validity as an independent disorder.

Helen Tager-Flusberg is director of Research on Autism and Developmental Disorders at Boston University.


1: Ozonoff S. J. Child Psychol. Psychiatry 53,1092-1094 (2012) PubMed

2: Skuse D.H. J. Am. Acad. Child Adolesc. Psychiatry 51, 344-346 (2012) PubMed

3: Tanguay P.E. Am. J. Psychiatry 168, 1142-1144 (2011) PubMed