I was looking into the recent SLT literature for stroke recently, and my very learned friend (who still works in academia) pointed me towards a new RCT about SLT for aphasia and dysarthria in the first four months after stroke. ”Hurrah”, I thought, “maybe the medics will start taking us seriously when we can point at our own set of RCTs”.
Here it is, freely available in full text:
The authors recruited people with aphasia or dysarthria following stroke, starting treatment about 17 days after their stroke. Half the group got best practice but realistic speech and language therapy over the next four months and half got visits from untrained visitors who did unstructured daily activities, mostly conversation. They compared outcomes at around 6 monthsprimarily by looking at functional communication measured on the TOMs activity scale.
Much to my horror, they found no difference in functional communication outcomes between the two groups.
That’s pretty stunning when you consider that as therapists, we would mostly love to be able to offer communication therapy a couple of times a week with as much impairment work as these participants received.
So what now?
Before we get too upset, it’s worth looking more closely at the study.
I salute and applaud the authors for trying to do the impossible, but there are big problems with RCTs in Speech and Language Therapy as a general service.
- They lumped dysarthria in with aphasia (though their wording suggests they didn’t mean to – ooops). These are two very different disorders, and even within aphasia there are many different subtypes. It’s reasonable to expect that these would have different prognoses for recovery and different responses to treatment.
- The therapy they provided was very mixed. Perhaps this is a result of trying to treat all the different impairments, or perhaps it is simply a reflection of what we do as therapists, but it makes it hard to say what you are testing.
- The volunteers also all did different things. Maybe they were doing something that was effective, but we don’t know what.
- They tried to use functional outcome measures. I agree with this in theory – no point in improving someone’s ability to point at something and say ‘giraffe’ unless they’re something they were keen on being able to do. However, when you use an 11 point scale that is rated by therapists watching you communicate with an unfamiliar listener, it’s not going to be very sensitive. There may have been improvements that were not detected.
This article annoyed my friend so much that she wrote a rapid response and concluded that “this is an issue that can only be addressed by targeted studies that use sensitive outcome measures for well-defined therapy approaches for specific communication disorders.”
As a therapist, why should I care?
It’s hard to know what I should take away from this study. I do think there are significant problems with it, but I’m not sure those problems completely explain away the fact that there was no difference between therapy and an attention control.
One thing we can know for sure is that the people who commission speech and language therapy services will read studies like this and we need to know how to respond.
I will certainly be aiming to make my therapy more functional and worrying less about not doing ‘real’ impairment level therapy unless I’m convinced that there’s evidence for it.
It also puts a certain amount of urgency behind the gradual increase in focus on outcome measures and outcome based commissioning. If we don’t have research that says we’re effective then we either need to be able to show it at the level of our service or change what we do.