This interview provides a look “Behind the Science” of the AJSLP article A Randomized Trial of 12-Week Interventions for the Treatment of Developmental Phonological Disorder in Francophone Children.
Designing the Study
The way that we think about speech disorders has changed quite a lot in the last few decades. Initially, I think, in our profession we thought about speech sound disorders in terms of the surface characteristics of the children’s speech. If they are not speaking very clearly, when we teach them, we need to go in and teach them how to articulate sounds more clearly. There has been a lot of focus on the surface characteristics of the child’s speech. More recently, we’ve come to understand that there’s a lot more going on down underneath your iceberg, and a lot of research on what we call levels of representation.
I had done some modeling of how all these levels of representation were linked together: the children’s knowledge of vocabulary, their perception of speech sounds, their phonological awareness, and their articulatory accuracy — their ability to say speech sounds accurately — and looking at how that related over time to their ability to learn to read. I had this model of how these different levels of representation related to each other, but modeling’s a way to develop hypotheses about causal relationships — you can’t exactly make claims about causal relationships. So I had wanted to do some intervention trials so that I could be more clear about what the causal relationships were.
Answering Pragmatic Clinical Questions
When I got to the point to where I said, “Okay I’m ready to do some intervention trials to test these hypotheses more clearly,” at the same time the local hospital came and they said, “We’re a bit distressed, we have to ration our services quite severely and go back to only six weeks of intervention.” There is no clear research on this, there are no randomized controlled trials or anything, but there is a lot of descriptive research, and that research suggests you can’t get a very good outcome in only six weeks. You might be able to do some good, but you’re not going to be able to have really measurable outcomes in terms of speech intelligibility or functional outcomes. So they said, “What should we do?”
Françoise had just started in my lab as a doctoral student, but she had actually been working at the same hospital as the speech language pathologists for quite a few years, and she had been treating their Francophone patients. So she had a lot of questions about how to do speech therapy better, but also how to treat the French-speaking children because there wasn’t a lot of research on that, and they were treating them as if they were English-speaking children — not a good idea. So she had a lot of questions, too.
We had designed an intervention for them that was actually 12 weeks long, but the components were very low cost. So we thought, you can do this for almost as low cost as a 6-week intervention with a speech pathologist. So we designed this intervention so it would solve the problem for the hospital: How can you provide a lower- cost intervention when there are resource constraints, but also how can we study the relationship between all these variables — vocabulary, phonological awareness, perceptual knowledge, and articulation learning — in these kids. And we’ll get the theory piece and the more pragmatic questions all in one study. And — Do it in French! Which was huge, a lot of questions all at one time.
French Is Not English
In retrospect, we should not have been surprised by this, but the biggest surprise for us was that French is not English.
We had to develop all of the outcome measures for the research ourselves. The errors that children make in French are very different from the errors that children make in English. The whole time that we were doing the study we were always just running to keep up. Because we would start making a test for phonological awareness or something, then we’re like, “Oops, that’s wrong, we have to start again.” Because, ridiculously — we know at an intellectual level that French is not English. But we didn’t really quite understand the implications of that until we started working with those children. People are always saying, “You can’t just translate things from English to Spanish” — or whatever language you’re working in. But oh man, is that ever true.
Testing Combinations of Intervention Approaches
So, what we had done with this study is we mixed and matched these different components.
Every child received an intervention that on the surface looked kind of the same. They received six weeks of individual therapy. Then in the second six weeks — this was the low-cost part — the children went for phonological awareness intervention in small groups, which is something that’s very low cost because almost anybody can do phonological awareness therapy. You could have high school students do it as volunteers or speech communication disorders assistants, anybody. So they did these little small group phonological awareness intervention. And while the kids were doing that, we brought the parents in as groups, and we taught them to do an intervention at home.
So we had these three components. But the components had different flavors. The individual intervention could be targeting articulation accuracy in the standard way — very standard, traditional articulation therapy — which we called our output- oriented intervention.
Or it was focused on listening to speech — a very input-oriented intervention. In this intervention, the children were presented with a lot of high- quality speech models by listening to stories that had the sounds that they were being taught. Or they would listen to speech and make judgments about the speech. So the clinician would present correct and incorrect versions of the word, and the child would have to say if it was the right one or the wrong one. They wouldn’t have to say that, but there would be little games that would do different things if the word was said correctly or incorrectly. And then, at some point towards the end, like the fifth or sixth session, if the child would choose to say some of the words, then we’d say, okay, we’ll do some minimal pairs production therapy and we would do a little bit of production therapy. But there was almost no speech practice, it was nearly all through the listening modality. The parents would watch these interventions, so they would see the speech pathologists implementing them.
Then in the second 6 weeks, the parent intervention had two different flavors. In one case, we would teach the parents how to do standard, traditional articulation therapy. Or we would teach the parents how to read to their children at home, using what’s called a dialogic reading approach — read to the children, ask questions while you’re reading. So, we had these two different kinds of parent interventions.
And then we mixed and matched these all up — so that we ended up with four different types of interventions. Two of them are coherent, in one case the parent is teaching their child how to articulate sounds accurately, and that is matched up with what the speech therapist is doing. So you have a nice, coherent progression of therapy. In the other case, the dialogic reading parent intervention is nicely coherently paired with that listening intervention that the speech therapist was doing. Because they are all very similar kinds of procedures.
Then we had two possibilities where they are mixed up, so that they’re not coherent.
Findings and Final Thoughts
So we found out a few things that are important. All of the interventions were more effective than having no intervention. Because we had this other waitlist control group. So that’s good. If you do therapy with kids, they get better.
What I think people might find surprising is that the children who had the intervention that was nearly all listening made just as much gain for articulation accuracy as did the children whose intervention was totally 100% of the time focused on articulation accuracy. And both of those groups with the coherent match ups of the parent intervention and the individual therapy, they also made equivalent gains for phonological awareness.
The two interventions in the middle that were all mixed and matched or not very coherent — they did not make as good gains.
So, I think there’s some important messages there. In some ways our results are similar to some things that other people have found. So, Anne Hesketh in the United Kingdom, she also found that an intervention that is focused on kind of the metalinguistic aspects of speech can be as effective as your traditional articulation therapy.
When speech pathologists see this, they say, “Well, this intervention works, that intervention works, that intervention works. It doesn’t matter what I do. I can do any therapy, it’s all going to be good. I can just pick whatever I’m comfortable with, whatever the parent is comfortable with, it doesn’t matter what I do, it’s all going to work.”
I think that that’s actually the wrong message. It does make a difference what you do. Because what we found is that how we combine these components mattered. I think the reason that it matters is that intensity matters. When we’re helping children learn how to speak clearly, there are all these different levels of representation that are involved. There’s a perceptual level, there’s the lexical level, there’s the articulatory level, and so on. And you have to tackle at least one of these levels with enough intensity that the child learns. And if they get learning at one of those levels, you will get generalization to the other levels. But you have to at least get learning at one level, and you have to have enough intensity at that level to get that learning. And that’s why it was so important for our interventions to be coherent — that the speech therapist and the parent were working together and tackling the same levels of representation.
Which is not the same as somebody saying, “Well, I can be eclectic, I can do anything. I can pick this therapy or that, or this procedure or that.” Because actually the decisions that you make and the way that you choose your procedures, and the way that you combine them together makes a difference.
It really makes a difference what you do. You could choose procedures, combine them together in a coherent way, and have a good outcome. You could choose procedures, combine them together in a non-coherent way, implement them badly, etc. and not have a good outcome. And I think this is actually good message for speech pathologists: We’re professionals, we’re very highly trained. It makes a difference what we do.