Labels for unexplained language difficulties in children: We need to talk

The view from the Tower of Babel

This week saw the publication of a special issue of the International Journal of Language and Communication Disorders, focusing on labels for children with unexplained language difficulties. Two target articles, one by Sheena Reilly and colleagues, and one by me, are accompanied by an editorial by Susan Ebbels, twenty commentaries, and a final paper where Sheena and I join forces with Bruce Tomblin to try to synthesise the different viewpoints. These articles are free for anyone to access.

Terminological battles are often boring and seldom come to any consensus, so why are we putting time into this thorny issue? Quite simply, because it really matters. As we argue in the articles, having a label affects how a children are perceived, what help they are offered, and how seriously their problems are taken. 'Specific Language Impairment' has very poor name recognition compared to dyslexia and autism, despite being at least as common. Furthermore, unless we can agree on some common language, it's difficult to make progress in research, and to discover, for instance, the underlying causes of language difficulties, how common they are in different parts of the world, or what interventions work.

I was first confronted with the full extent of the problem when I tried to analyse the amount of research and research funding associated with different developmental disorders (Bishop, 2010). There are other conditions, notably autism and dyslexia, where there is plenty of debate about diagnostic criteria, or even about whether the condition exists. But even so, the terminology is reasonably consistent. For children's language difficulties, this is not the case - they can be described as cases of language difficulty, disorder, impairment, disability, needs or delay, with various prefixes such as 'developmental', 'specific' or 'primary'. Some researchers will use such labels with precise meanings, often excluding children who have co-existing conditions, whereas others use them more descriptively. This made it extremely difficult to do a sensible internet search to estimate the amount of research funding associated with children's language difficulties.  

The confusion over labels has, I think, also contributed to the lack of public recognition of language difficulties in children. A couple of years ago, I joined together with Courtenay Norbury, Maggie Snowling, Gina Conti-Ramsden and Becky Clark with the goal of remedying this situation. We started a campaign for Raising Awareness of Language Learning Impairments (RALLI) (Bishop et al., 2012), and set up a YouTube channel to provide basic information. We spent some time debating what terminology to use: "Language learning impairment" was our preferred choice, but many of our videos talk of Specific Language Impairment, simply because that is a more familiar label. The lack of an agreed label proved a real stumbling block for our attempts at public engagement, and we decided that, as well as producing videos, one of our goals would be to get the terminology issue discussed more widely, in the hope of achieving some consensus. It was a very happy coincidence that Sheena Reilly and colleagues were crystallizing their own position on this question in an article in IJLDC, and that they, and the Editors, were willing to include my article, and the commentaries of other RALLI founders, in the published debate.

One thing that came across when reading commentaries on our articles was the disconnect between research and practice. One point on which I agree with Sheena and colleagues is that there is no justification for drawing a distinction between children whose language problems are comparable with below average nonverbal ability, and those who have a mismatch between good nonverbal skills and low language. Research has failed to find any difference between children with uneven or even nonverbal-verbal profiles in terms of responsiveness to intervention or underlying causes. Such a distinction is, however, widely used in educational and clinical settings to decide which children gain access to extra support in school.  Another issue raised by the Reilly et al paper is whether it is logical to use other exclusionary criteria, and to distinguish, for instance, between children who do and don't have autistic features in association with a language problem.  

As Susan Ebbels noted in her editorial, in everyday settings "diagnostic labels and criteria were being used creatively in disputes over access to services both by those seeking to obtain services for children (often parents and their lawyers) who could be accused of ‘diagnostic shopping’ and also by those seeking to deny services (often due to financial constraints) who may use particularly restrictive criteria in order to reduce the number of children qualifying for services". 

We can't afford to ignore this confused situation any longer. The time has come to have a wider debate on these issues, with the aim of reaching a consensus about how terms are used. The Royal College of Speech and Language Therapists has set up a moderated discussion forum where people can give their views on the best way forward. Please do consider adding your voice: it is important that all those affected by this issue have a say, whether you are a speech-language therapist/pathologist, psychologist, teacher, health professional, legal expert, policymaker, a parent of a child with language difficulties, or someone who has experienced language difficulties. We'd also love to hear from those outside the UK - whether English-speaking or not. You can access the discussion forum here.

Finally, to raise awareness of this debate, during the week of 24th-31st August I will be taking over  the @WeSpeechies Twitter handle as guest curator. On Tuesday 26th at 8.a.m. BST there will be a live twitter debate on this topic. Feel free to join in, even if you aren't a regular tweeter.

References
Bishop, D. (2010). Which Neurodevelopmental Disorders Get Researched and Why? PLoS ONE, 5 (11) DOI: 10.1371/journal.pone.0015112  
Bishop, D., Clark, B., Conti-Ramsden, G., Norbury, C., & Snowling, M. (2012). RALLI: An internet campaign for raising awareness of language learning impairments Child Language Teaching and Therapy, 28 (3), 259-262 DOI: 10.1177/0265659012459467

Slides on this topic are available here.

Addendum Friday 29th August 2014

We've had a great week of interactions on Twitter. A transcript for the week is available here.

I'll look through this and aim to organise the material in due course, but meanwhile would encourage anyone who is interested to continue the discussion on Twitter. I'm appending below some tweets that I generated throughout the week to generate debate.

As noted above, the chat links in to a special issue of the Internat. J Lang. Comm Dis which is free to access here http://t.co/ncTUaYvyoI.  NB it is not all that obvious but there are 10 commentaries after each target article.

If you want to join the discussion on Twitter, feel free to comment at any time, but, please include the #WeSpeechies hashtag, so we can aggregate comments easily. Also if your comment relates to a numbered question, please add Q1, etc so we can relate them.

Monday started with my attempt to summarise each of the  twenty commentaries in a Tweet-length message.

Summaries from commentaries

Paediatricn Gillian Baird: ICD &DSM classifications talk of 'language disorder'; implies distinct from normal variation.  Disorder’ used for conditions without obvious aetiology; functional effect described separately in ICFDH.

Lauchlan/Boyle, ed psych view. Must ask: ‘Will label change the child's life for the better? Aetiology often irrelevant

Bellair et al: community SALTs. No one label works for both research & clinical. SLI has problems but we can manage them.

Mabel Rice: "SLI has yet to receive widespread adoption in clinical practice, in spite of the great need for it." critical of DSM5: excluded "well-researched category of SLI", included SCD, "with a minimal research base"

Kate Taylor SLP. SLI underidentified. Changing the term won't resolve the issue, which is one of measurement rather than label.

Conti-Ramsden: Any Consensus Panel on terminology must be international and include voices from different languages,

Hansson et al: ICD10 labels don't map on to use by researchers in Sweden . : Sweden: phonological & grammatical difficulties seen as part of language impairment. Soc comm probs separate

Clark & Carter: Survey:Scottish SALTs unclear re terms & diagnostic criteria. Move from exclusionary to inclusionary criteria.

Hüneke & Lascelles http://t.co/9rVKJzoBZV. Concern that watering down terminology will mean kids lose scarce resources. Prefer medical term 'developmental dysphasia' that gets problems taken seriously

Strudwick/Bauer http://t.co/GSY5Xwz283 Concern that labels don't capture comorbidities; most ch with 'SLI' have other problems

Michael Rutter, psychiatrist "both clinical & research classifications needed but they require a different approach"

Rutter: Specific’ implies ‘pure’ language impairment; "not supported by any of the available evidence"

Larry Leonard: Many researchers already use broader definition of SLI: do not use term to mean children have a pure profile. communicatn with the public/other disciplines will be even harder if we adopt generic label ‘language impairment.

Snowling: DSM5 treats Communication Disorders separately from Specific Learning Disorders, yet they often co-occur

Aoife Gallagher,SALT; ethical issue:"who owns diagnosis once it has been given.. who ultimately has the right to take it away"

Andrew Whitehouse: ‘SLI’ provides neat criteria for researchers but label hides behavioural & aetiological heterogeneity

Dockrell/Lindsay Educational perspective re SLI is missing yet day-to-day support of learning/development provided by teachers. in England ‘speech, language & communication needs’ (SLCN) indicates primary need is with language & communication

Grist & Hartshorne: http://t.co/QKeQbQFsdy Children & young people we work with rarely describe selves as having SLI or SLCN

Norbury @lilaccourt Relaxing diag criteria will increase demand for services.SALTs shld focus on severe & persistent impairmts

Parsons et al @wordaware Shockwaves through SALT profession if nonverbal IQ criteria and delay/disorder distinction removed .Use of marketing approaches to development of a new term, including consultation with parents & young people.

Wright: legal perspective Much time spent in tribunal appeals arguing re labels: eg is it delay or disorder, is it specific?

Questions for debate

On Tuesday we had a live twitter chat with four question topics, and later in the week, I added further numbered question. Here is the total list – we'd love to hear your thoughts on any or all of these:

Q1 What is your view on use of the diagnostic label SLI? Does it reflect a medical model and is this appropriate.

Q2 is What are appropriate criteria for identifying children's language problems

Q3; Should IQ, ASD features, hearing loss determine whether language-impaired children can access services?

Q4 What terminology is most appropriate for children who have unexplained language problems?

Q5 ICD11 will use'Developmental Language Disorder' and DSM5 uses 'Language Disorder'. What do people think of these terms?

Q6 In research SLI still widely used but without requiring IQ discrepancy. Should we retain SLI but with this broader meaning, or is it just confusing?

Q7 In UK education, Speech, Language and Communication Needs (SLCN) is popular term. Is it used outside UK? Is it useful?

Q8 In UK clinical practice, distinction between language 'delay' & 'disorder' is used, but it has no research support.  Where does delay/disorder distinction come from? How defined?

Q9 Is there any support for a return to the more medical term 'developmental dysphasia'?

Q10. Reilly et al and several commentators suggest we drop 'Specific' and use the term 'Language Impairment' instead .What wld be advantages (e.g. avoids unfair exclusion) and disadvantages (e.g. too broad)?

Q11 What do people think of terms 'Language Learning Impairment' or 'Primary language impairment'? '

Q12 Do diagnostic labels actually help children and families?

Q13 Shld terminology/diagnostic criteria be responsibility of speechies, or shld other professions & families have a say? Assumptions/practices seem v. different in education/medicine/psychology vs speech-language therapy/pathology

Q14 In yr area, who does intervention with kids whose language problems are associated with autism?

Q15 Some  people take pride in identifying themselves as dyslexic. Does this ever happen for kids with language problems? If not, why not?

Q16 Has anyone encountered situation where child not offered intervention bcs language problems attributed to social deprivation?

Q17 Insurance considerations seldom important in UK, but affect label use elsewhere. Do US insurers just require DSM?

My thoughts on the dyslexia debate

©CartoonStock.com

During February, there was widespread media coverage of a forthcoming book by Julian Elliott and Elena Grigorenko called The Dyslexia Debate. I've seen an advance copy of the book, whose central message is to argue that the construct of dyslexia lacks coherence. Quite simply, dyslexia does not constitute a natural category, in terms of cognitive profile, neurobiology or genetics.

The authors' arguments are backed by a large body of research: people have tried over many years to find something distinctive about dyslexia, without success. Some children are good at reading and others are not, but it's arbitrary where you place a cutoff to specify that a child has a problem. There's a popular belief that you can identify dyslexics in terms of a particular ability profile, and that affected children have a particular kind of brain organisation that makes them do things like reverse letters (b vs d) or have left-right confusion. In fact, those types of problem are common in typically-developing children at early stages of learning to read and appear to be as much a symptom as a cause of reading problems. Researchers have found neurobiological and genetic correlates of developmental reading problems, but the effects tend to be small and inconsistent from person to person: you could not currently diagnose dyslexia on the basis of brain scans or genetic analysis. It is, of course, possible that one day we may hit upon a new diagnostic test that does clearly differentiate a dyslexic subgroup from other poor readers, but many of us in the field are dubious as to whether this will ever happen.
 
The first thing to get clear is that Elliott and Grigorenko are not denying the reality of children's reading problems. Their point is a much more specific one about the way we conceptualise reading difficulties and how this affects access to services in everyday life. Their concern is that "dyslexia" implies we are dealing with a specific medical syndrome. Their view is that no such syndrome exists and it is not helpful to behave as if it does. How should we respond to this? I think we need to distinguish three questions:

1. Should we identify those in need of extra help?

Children vary in the ease with which they learn to read. Some need only the briefest exposure to books to crack the code; others struggle for years despite skilled help from expert teachers.
I think most people who have spent time with poor readers (and I would include Elliott and Grigorenko among these) would conclude that the answer to question #1 is yes. It doesn't matter whether those in the latter group have a distinct medical syndrome or not: It is up to us to ensure that all get the best teaching.

2. How should we identify those in need of extra help?

Elliott wrote a piece in the Times Higher Education where he argued that dyslexia diagnoses in Universities were skyrocketing, and that some people were unfairly exploiting the system in order to get accommodations such as a laptop computer and extra time in exams. To my mind, the problem here is less to do with the "dyslexia" label, and more to do with the haphazard way in which individuals are identified, and the lack of consistent criteria for determining who needs extra help. Like Elliott, I think it is entirely right that we should make accommodations for students who have serious difficulties in processing written information at speed. However, as he highlights, the current system is based on an unsustainable idea that "dyslexia" is a distinct disorder that can be reliably identified, and which is often diagnosed on the basis of supposed markers of dyslexia that have no scientific basis. So the current system is both invalid and unfair. Instead, it would be sensible to settle on consistent criteria for allocating extra help to students who are struggling, and to ensure that extra resources are directed to those who are most needy. As Castles and colleagues have noted, there are guidelines that can be used to identify those with severe and persistent problems, but they are not well-known or widely applied.

3. What terminology should we use to refer to those we identify?

So can we just agree that we need to find consistent ways of identifying poor readers and do away with the term "dyslexia"? While this might seem a logical response to the evidence, I think we should not underestimate the implications in practice. On the positive side, we'd get rid of the idea that we're dealing with a special condition that forms a distinct syndrome. Since few scientists would attempt to defend that notion, this would be a good thing. But we should also be aware of negative consequences.

Those commenting on the dyslexia debate so far have talked about it as if it is a particular issue relating to literacy difficulties, but in fact it's just one instance of a much more pervasive problem.  Other neurodevelopmental disorders such as autism spectrum disorder, specific language impairment, attention deficit hyperactivity disorder, developmental dyspraxia and dyscalculia are all beset by the same issues: there is no diagnostic biomarker, the condition is defined purely in terms of behaviour, different disorders overlap and there's no clear boundary between disorder and normality.
 
Similar issues have been much discussed in relation to adult psychiatric disorders, which are also diagnosed in terms of behavioural features rather than biological tests. In a fascinating paper, Kendell and Jablensky (2003) came to the conclusion that the categories of schizophrenia and depression are massively problematic in terms of validity and reliability – that is to say, just like dyslexia, they don't constitute natural categories clearly demarcated from other disorders, and furthermore, people can't even agree on who merits these diagnoses. So should we just stop using the labels? Kendell and Jablensky considered this possibility but concluded it would be impossible to abandon terms like schizophrenia and depression, on the grounds that they have utility. These labels have been used for many years by practitioners to determine the most effective intervention, and by researchers interested in discovering the underlying causes and likely outcome of a disorder. Similarly, using the construct of "dyslexia" we have discovered much about the nature of the cognitive deficits that characterise many poor readers, about underlying causes, about outcomes, and about effectiveness of intervention. For instance, we know that genes play a part in determining who is a poor reader, and that many children who have poor literacy skills also have subtle problems with oral language.
 
This argument, though, is not really watertight. We may congratulate ourselves on what we have learned, but on the other hand, it could be argued that there are also barriers to progress that arise from continued use of imprecise terms. It's clear to anyone who knows the research literature that findings can vary from study to study and from child to child within a study. This does not necessarily invalidate the research – it's rare to obtain perfect consistency of findings even within mainstream medicine – but it does make many people wonder whether we might obtain clearer results if we took a different approach. But then we have to consider what alternative approach would be better.

I suggested a few years ago that it might be helpful to treat neurodevelopmental disorders differently, as multidimensional composites, rather than regarding problems with reading, language, arithmetic, attention, motor skills and social behaviour as separate conditions. However, I did not really expect anyone to embrace this idea, as it would be too radical a change, and we are too wedded to current terminology.

Here too, comparisons with psychiatry are interesting. Last year, Tom Insel, director of the US National Institute of Mental Health, ruffled feathers by stating that his organisation would be reorienting its research away from traditional psychiatric diagnostic categories, to develop Research Domain Criteria, i.e. "new ways of classifying mental disorders based on dimensions of observable behavior and neurobiological measures." Yet the domains that are proposed seem to me just as arbitrary as the original diagnostic categories, and the associations between genetic, neurobiological and behavioural measures are mostly weak and poorly understood. So although Insel's vision might seem a rational way of trying to make sense of psychiatric disorders, it is years away from being clinically applicable – as he is the first to admit.  Even though multivariate, dimensional classification seems more logical, our current categories of autism, schizophrenia and dyslexia, though imperfect, may be as good as we can manage in terms of utility in day-to-day clinical practice.

Perhaps the strongest arguments in favour of retention of a term like "dyslexia" come not from science but from public perception. Like it or not, "dyslexia" has been around for over 100 years. In that time, a range of organisations have sprung up to help people with this diagnosis. Some of the most passionate defences of the dyslexia label come from those who have built up a sense of identity around this condition, and who feel they benefit from being part of a community that can offer information and support – see, for instance, this comment by the International Dyslexia Association to the suggestion that "dyslexia" be removed from the DSM5.

One could, of course, argue, that we shouldn't stick with a label just because it has always been there – if we were to adopt that line of argument, we'd still be talking about "maladjusted" and "educationally subnormal" children. But it's clear that many of those diagnosed with dyslexia do see this label as positive. In particular many people worry that if they were to simply switch to a more neutral, less medical term, such as "poor readers", this could trivialise reading problems, and lead people to assume that the difficulties are just caused by poor teaching. Furthermore, legal entitlement to special help under disability legislation could disappear. This, I think, is a key part of the problem, which can get overlooked when just focusing on the scientific evidence: what you call a condition determines two things: how seriously people take it, and where they place blame for the difficulties and responsibility for doing something about it.

To illustrate my point, see this recent piece in the Daily Mail by Peter Hitchens, which appeared under the headline: "Dyslexia is NOT a disease. It is an excuse for bad teachers". This displays a remarkably simplistic world view in which a poor reader either has a "disease", in which case they are blameless victims of an external force, or else it is someone's fault – in this case lacklustre teachers.

In his triumphalist piece against the "pseudoscience and quackery" of dyslexia, Peter Hitchens achieves exactly the opposite of what he intends. This is because he demonstrates one negative consequence of removing the label, which is that many people will no longer think that children who struggle to read need any kind of special help. Instead, we'll be told that "What they need, what we all need, is proper old-fashioned teaching."

A rather more sophisticated version of the same argument was given in the Green Paper that introduced the Government's proposed revision to legislation for Special Educational Needs (SEN) (see: my blogpost on this). There it was stated that too many children were being over-identified with SEN: “Previous measures of school performance created perverse incentives to over-identify children as having SEN. There is compelling evidence that these labels of SEN have perpetuated a culture of low expectations and have not led to the right support being put in place.” (point 22).

We really need to escape this polarised view of children's problems being caused either by a medical disease or by poor teaching. Yes, some children's reading may be held back because their teachers either don't know about or reject evidence-based methods of teaching, but it is seldom black and white, and some children fail despite intensive, high-quality teaching.

My concern is that those holding the purse-strings have a strong incentive to blame all problems on bad teaching or bad parenting, as it absolves them of any responsibility to do anything about them. We need to recognise that for most children, the causal influences are likely to be complex and may involve both constitutional factors and aspects of home and school environment. Unfortunately, most people don't seem able to deal with this complexity, and the language we use determines how problems are viewed. At present we are between a rock and a hard place. The rock is the term "dyslexia", which has inaccurate connotations of a distinct neurobiological syndrome. The hard place is a term like "poor readers" which leads people to think we are dealing with a trivial problem caused by bad teaching.

As Allen Frances argued in the case of psychiatry, we need to resist a growing tendency to use medical labels for what is essentially normal behaviour. However, he wisely notes that this should not blind us to the reality that there are people with problems that are severe, clearcut, and unlikely to go away on their own.  In the current debate, several commentators have made this point and have added that it doesn't really matter what we call them; the more important issue is to ensure affected individuals get appropriate help. But I'd suggest it does matter, because the label we use does much more than just identify a subset of people: it carries connotations of causation, blame and responsibility. While I can see all the disadvantages of the dyslexia label outlined by Elliott and Grigorenko, I think it will survive into the future because it provides many people with a positive view of their difficulties which also helps them get taken seriously. For that reason, I think we may find it easier to work with the label and try to ensure it is used in a consistent and meaningful way, rather than to argue for its abolition.

Reference
Kendell, R., & Jablensky, A. (2003). Distinguishing between the validity and utility of psychiatric diagnoses American Journal of Psychiatry, 160 (1) DOI: 10.1176/appi.ajp.160.1.4


This article (Figshare version) can be cited as:
Bishop, Dorothy V M (2014): My thoughts on the dyslexia debate. figshare
http://dx.doi.org/10.6084/m9.figshare.1030404