Is it important to start stuttering treatment early?
I’v read many places that it’s important to start stuttering treatment early – as early as possible after onset. My problem is that all the references to evidence meant to support this does not tell me that the success rate/improvevement is higher when starting early, but instead things like:
1) Early treatment (like the Lidcombe program) helps
2) There is a recovery dip after 10-15 months (Yairi et al, who did not induce any training, but just observed stutterers), meaning that if you haven’t improved much after about 15 months after stuttering onset you are less likely to recover.
3) There are different treatments for young and more adult stutterers (one study with four subjects claims – without really studying it – that it takes less time to get an effect with younger children, but this is quite weak evidence)
4) No reason to believe that it is harmful to start with an early treatment (which means that improving early is better, since it reduces a handicap)
I suspect that the reason for believing in the importance of early treatment is the plasticity of the brain, which is better the younger you are. But if increased plasticity is essential for stuttering treatment, I wonder why the few studies (three in total) I find on this question does not find any effect of earlier treatment.
So, am I missing something here – or is it just not documented (rather the opposite) that early treatment is important for the success of stuttering treatment?
NB: If it really is a big decrease in success rate when waiting, what is currently done is really bad and there should be changes in the prioritisation of SLP resources. Just to avoid misunderstandings, I wouldn’t recommend “wait and see” for you young children, since there are other negative potential implications of a childhood stutter.
Probably the best way to think about this is from a common sense perspective. With randomised controlled evidence showing favourable odds ratios for early intervention, and with evidence of negative consequences of stuttering beginning during the school years, early intervention is important. Common sense will also suggest that treatment soon after onset is not the best use of health care resources, because there is some chance of natural recovery shortly after onset.
Thanks for the answer, but what I’m looking for is the empirical evidence showing favourable odds ratios for early interventions compared with later interventions. Sorry, if that was unclear from the question. If the success rate is much higher when starting very early, e.g., just after onset, I think that the use of the resources would be defendable, even when most would recover without treatment. All this hangs on the effect from early compared to later treatment.
I think it is easy to agree with you on that something unpleasant (with possible negative consequences) should be removed/reduced as early as possible, but what I read many places is that it is important due to the higher rate of success/lower effort to reach the success. Is that documented?
Magne, this is a great question and a very important one. I do agree with Mark that with children very near the onset of stuttering who have no risk factors for continued stuttering that intervention is not a good use of time for parents, children, and the speech-language pathologist. To my knowledge, there are no empirical studies on the efficacy of early intervention (say, within 6-12 months of onset) vs. later intervention (say, 18 months and beyond), and this is frustrating for clinicians and families. Some parents come to me 2 years or so post-onset, feeling guilty that they have not come in sooner. I tell them that there is no empirical evidence that coming in earlier will make therapy goals easier to achieve.
Yes, younger children have more malleable brains and motor skills, and this may make earlier intervention more successful, but we do not have the evidence for that. I worked in the labs of Ehud Yairi, and saw many children recover from stuttering with no formal treatment. Certainly development and natural recovery have a role to play in treatment programs for preschool children.
Dear “magne,”
Good question. I am curious who you are. Wanna identify yourself?
Mark has lots and lots of evidence on the Lidcombe Program for treating preschool stutterers. I won’t presume to know more than an objective observer’s knowledge of the program, but I have read that it works better for children who are closer to 5 years of age than 3 years of age. Perhaps I am mis-informed, but I did attend a seminar that Mark presented comparing the Lidcombe Program to spontaneous recovery. If what I read/heard was true, this might shed light on Mark’s suggestion that common sense might motivate one to wait at least for some months for spontaneous recovery before beginning to treat a 2-3 year old who has begun to stutter.
That certainly makes sense, but I would agree that your question of which is better, immediate or delayed treatment, has not been sufficiently answered by the extant research. Obviously, it would be difficult research to carry out if a standard, replicable treatment cannot be applied to beginning versus somewhat established stuttering due primarily to the age and maturity of the children involved.
I will tell you that I advocate beginning treatment as soon as possible, (but not necessarily within the first few days after stuttering is noticed). I know full well that typical preschool-onset stuttering lasts weeks or months and that spontaneous recovery (whatever that is?) can be assumed to be responsible. Nevertheless, early treatment in the cases where spontaneous recovery does NOT occur can have two positive effects that later treatment might not have. First, the sooner stuttering can be eliminated (or greatly reduced), the less likely strong habitual patterns surrounding the stuttering can be mitigated. Second, the sooner such improvements can occur the less likely the child is to begin to see himself/herself as somehow incapable of normal speech, develop negative emotions regarding the stuttering, etc.
You are right that early treatment engenders little or no risk to the child of making it worse, as Wendell Johnson would have predicted. Given that, the main downside to very early treatment is that there is cost and effort involved to the parents and family. In my judgement, the potential benefit (easier, often less involved treatment) far outweighs the costs of waiting (potentially more complex, emotionally-laden stuttering).
Ken St. Louis
No problem identifying myself. I’m Magne Jørgensen from Norway, very interested in stuttering research, but have no formal background in it, have stuttered since I was about 4-5 years old, used “the McGuire-treatment” (regulated breathing-like + other evidence-based elements) to achieve a 99% recover from stuttering. Not that interesting 🙂
Your answers to my questions, on the other hand, are very interesting – especially since I know your extensive expertise in the field. As I understand you, you find (like me) that there are no documented evidence of a difference in effect of early vs delayed treatment. This does not mean that there are none, but more that it has not been properly studied.
It is, I think, interesting that the – very important – question of the timing of the stuttering treatment (even those with well-documented effect) is so open to debate and that there are disagreements based on different non-empirical argumentations. In theory, it could be the case that starting the treatment just after onset could improve the stuttering treatment a lot. I know that this is hardly likely to be true, especially if the “white matter” theories of why we stutter is valid, but nevertheless something that should be studied.
I do not fully understand why designing studies to address this topic should be that difficult or unethical.
Two examples outlining my (possibly naive) ideas about possibles studies:
1) Experimental studies: Very early treatment. Get funding to treat a random sample of children (who otherwise would not get any, or just very sporadic treatment) as early as possible after onset. Given previous studies I hope that we know enough about spontaneous recovery rates to meaningfully adjust for that factor. Start treatment of the other children later (after one year, two years, …) with (one of) the most effective treatment for that age group. I see that this can be perceived as unethical, but given that not all children are treated today (and the other will hopefully get the normal, usually not very effective – once-a-week visit to the SLP at school) I believe that such experiments would be defendable and not different from the, for example, the “wait group” in “Lattermann, Christina, Harald A. Euler, and Katrin Neumann. “A randomized control trial to investigate the impact of the Lidcombe Program on early stuttering in German-speaking preschoolers.” Journal of Fluency Disorders 33.1 (2008): 52-65.”
2) Observational studies: There will be a natural variation of when stutterers start their treatments. If there are no indications of that those who start earlier get better than those who start later, this would suggest no strong effect from early treatment. (The lack of control group makes causal claims more complex, but this type of study is possible to do with many data points – inexpensive data collection – and could give an indication when adjusting for factors related to severity of stuttering, dividing into subgroups of stutterers etc..)
By the way, the Wendell Johnson-ideas about inducing stuttering made my mother feel guilty about my stuttering for many years, until I discovered that his “study” on this was of extreme low quality. Hard to debunk myths once spread …
Wonderful discussion, and interesting suggestions above, Magne! I just want to add two more aspects into this discussion/dialogue. There seems to be cultural differences between countries regarding treatment towards preschoolers. One aspect is related to the content in treatment and the second aspect is related to what we define as “early intervention”. There might be a bigger variability in treatment towards preschoolers in Europe compared with for example Australia (correct me if I am wrong, Mark). Many approaches include supervisions to parents. In my practice I experience that the more informed the parents are of stuttering, possible risk factores and different choices for treatment, the more optimal their behaviour is towards the child and his/her stuttering.In this sense, I recoommend early intervention too, even though the evidence might only be practice based. Hilda Sønsterud
Interesting discussion. Yes, studies comparing early vs late treatment would provide the clearest evidence, but I think there would be ethical problems doing them. We have pretty clear evidence that delaying the Lidcombe to after age 7 or 8 decreases its effectiveness (longer time in treatment with outcomes not as consistently good as for children who start in preschool). We also can see the results of treatment programs for school-aged children, like the ELU and GILCU. Especially for the ELU, the pattern is more therapy sessions to achieve poorer results that what is possible with younger, more shape-able brains.
Indeed, it would not be ethical to compare and early intervention and delayed intervention treatment group in a trial. There have been no randomised trials for school age children, but it certainly is correct that the nonrandomised evidence for school age children suggests that treatment is far less efficacious and effective at that age.
Could you direct me to any published evidence on this? This is the type of evidence I’v been looking for.
Non randomised outcome data for the Lidcombe Program with pre-schoolers show post-treatment stuttering rates of below 1.0 %SS. For example:
Rousseau, I., Packman, A., Onslow, M., Harrison, E., & Jones, M. (2007). An investigation of language and phonological development and the responsiveness of preschool age children to the Lidcombe Program. Journal of Communication Disorders, 40, 382–397.
Onslow, M., Andrews, C., & Lincoln, M. (1994). A control/experimental trial of an operant treatment for early stuttering. Journal of Speech and Hearing Research, 37, 1244–1259.
Miller, B., & Guitar, B. (2009). Long-term outcome of the Lidcombe Program for early stuttering intervention. American Journal of Speech-Language Pathology, 18, 42–49.
However, non randomised outcome data for the Lidcombe Program with school age children show post-treatment stuttering rates of above 1.0 %SS. For example:
Lincoln, M., Onslow, M., Lewis, C., & Wilson, L. (1996). A clinical trial of an operant treatment for school-age children who stutter. American Journal of Speech-Language Pathology, 5, 73–85.
Koushik, S., Shenker, R., & Onslow, M. (2009). Follow-up of 6–10 year-old stuttering children after Lidcombe Program treatment: A Phase I trial. Journal of Fluency Disorders, 34, 279–290.
Additionally, Table 5 of this randomised controlled trial
Jones, M., Onslow, M., Packman, A., Williams, S., Ormond, T., Schwarz, I., Gebski, V. (2005). Randomised controlled trial of the Lidcombe programme of early stuttering intervention. British Medical Journal, 331, 659–661.
Shows no change in responsiveness by age during the pre-school years.
Interestingly, the data in the Kousik et al report cited above show a significant negative correlation between age and outcome, although for some unknown reason it is not reported in the paper.
Thank you very much for the references. I was able to read all, but two of them.
What I find is that all of the above studies report NO effect (or negative effect as in Koushik) from treatment age or from time from onset to treatment. This means, as I understand these results, that available evidence indicate that it is not important for the treatment effect (BUT, here are other negative effects of late treatement) whether the treatment is started at age 6 or at age 10.
To use these studies as evidence for importance of treatment in younger age vs. starting when youth/adult would require that we could compare the effect sizes (success measures) of younger <10 meaningfully with the effect sized for youths/adults.
Sorry to bother you with another request, but are there such comparisons (I'v found none, and there is always this isssue of selection bias/natural recovery rate more likely for those with less severe stuttering, which also requires less effort to improve)?
Here are my very brief summaries of the results:
Rousseau, I., Packman, A., Onslow, M., Harrison, E., & Jones, M. (2007). An investigation of language and phonological development and the responsiveness of preschool age children to the Lidcombe Program. Journal of Communication Disorders, 40, 382–397.
Result: Onset to treatment time NOT related to treatment time.
Jones, M., Onslow, M., Packman, A., Williams, S., Ormond, T., Schwarz, I., Gebski, V. (2005). Randomised controlled trial of the Lidcombe programme of early stuttering intervention. British Medical Journal, 331, 659–661.
Result: Showing no difference in effect size of Lidcombe between those 4 years old.
Lincoln, M., Onslow, M., Lewis, C., & Wilson, L. (1996). A clinical trial of an operant treatment for school-age children who stutter. American Journal of Speech-Language Pathology, 5, 73–85.
Results: Could not download this, but nothing in the abstract that indicate age effects.
Onslow, M., Andrews, C., & Lincoln, M. (1994). A control/experimental trial of an operant treatment for early stuttering. Journal of Speech and Hearing Research, 37, 1244–1259.
Results: Could not download this, but nothing in the abstract that indicate age effects.
Miller, B., & Guitar, B. (2009). Long-term outcome of the Lidcombe Program for early stuttering intervention. American Journal of Speech-Language Pathology, 18, 42–49.
Result: No age-related or stuttering-onset-to-treatment correlation to treatment effort (number of clinic visits required).
Koushik, S., Shenker, R., & Onslow, M. (2009). Follow-up of 6–10 year-old stuttering children after Lidcombe Program treatment: A Phase I trial. Journal of Fluency Disorders, 34, 279–290.
Result: Shows a strong NEGATIVE correlation (only 12 participants). Took more effort reach improvement and the improvement were smaller. (Even if taking the log of the improvement or the relative improvement). Not showing that starting early is important, rather the opposite.
Assume that a trial had as a goal to find out something essential (e.g. whether treatment very soon after onset is much more effective than later treatments) AND noone receives less treatment than otherwise (but a few receives more), whould that trial still be unethical?