Filter by Categories
Clinical Practice Research
Ethical Legal and Regulatory Considerations
Planning, Managing, and Publishing Research
Research Design and Method

Testing Theories: Using Intervention Research to Build a Theoretical Base for Evidence-based Practice

Nadine Martin

DOI: 10.1044/cred-int-bts-002

People think treatment research is just about finding a good method. It’s actually got more value — it can also be used to test theories. And that’s very important. The hardest part, I think is finding the interesting research question to begin with. It has to be of interest to you.

I must say, I came up with my question because I worked as a clinician for a number of years. It began with an observation. That individuals who have language impairments in aphasia also always have restricted verbal short-term memory spans. That spurred my investigation of why are these two related. How could these be related?

At that time, there were two theories about how language and short-term memory were related. One was that words that we process are just sort of processed and then held in some separate verbal short-term memory store. Then the other approach was that retaining information for short periods of time happened at all levels of word processing. So you hold onto its meaning for a short time, its sounds. Your brain has to hold onto these things while you’re preparing to say a word, or if you’re repeating a word — so the verbal short-term memory is really intrinsic to word processing. I was intrigued by this idea, and I said, “I think that may be why they are both impaired in aphasia.”

My theoretically-oriented testing began with trying to find evidence for that. There were two sources of evidence for that that I can talk about. One is developing associations between two cognitive abilities. So, one ability is word-processing. You need a good measurement of word-processing. Another is a good measurement of short-term memory span. Over a number of years, I collected data from about 50 or 60 people with aphasia, and I collected it on measures of word processing, picture naming, and word recognition — so, is this a word or not a word. Then each one also got measures of verbal span. Then what you do is you run regression analyses associating or correlating one measure with the other. The goal is to find out if there is a severity continuum — that is, you want to find an association between a person’s level of impairment of the word-processing and how short their verbal short-term memory span is. You would expect smaller spans with more severe impairments. This is called a severity continuum. We did find that association. So that’s one approach to finding that.

And then the second approach: We had the good fortune of evaluating an individual longitudinally. We began to test him very early on. He had a stroke after surgery for an aneurysm. We followed him for a year or two, and after about a year and a half, his word processing got better. There’s other elements that I’m not going to go into detail about, but his pattern of error in repeating words changed for the better, and his span got bigger. That’s just an example of how there’s another type of association. As he gets better, both of these abilities get better in parallel.

And we have other studies that we did that showed certain associations between these two. But associations are not enough to establish a causal relationship — that verbal short-term memory is causing the word-processing, or vice versa. It just simply says they’re associated. That really was part of our theory, that the language problem is actually holding onto or maintaining the verbal representations. So the short-term memory is what’s intrinsic to the word-processing itself.

So, another approach to testing this, and maybe establishing a more causal approach, is to actually try to treat the disorder. If we had a third construct that we said caused both these impairments — and that was the ability to hold onto or retain the activation of the semantic and phonological representations of words. So, to test that you have to actually treat the problem. If you say that that’s the problem, you have to treat that problem, and both should get better. Just the way the young man’s abilities got better as he recovered, so we hypothesized that was the ability that was getting better.

So that’s where we’re at right now. We have a couple of case studies where we’ve actually developed a treatment protocol — where we kind of turned the problem on its head. If someone can’t hold onto information for a very long time, that’s what we’re going to treat. We’re going to ask them to process words, but with increased memory load. It’s a very simple means of increasing memory load: We ask them to hear a word and then wait five or ten seconds and then repeat it. And they’re very bad at that, their performance goes down. But then with training, they get better. So, you can begin to demonstrate a causal relationship if what you say is the problem, and you’re treating that directly, and the symptoms get better, then you have better grounds for saying one causes the other.

And, one more thing I wanted to say, just about methodology. You can do this in a whole treatment study. But it’s probably better, first, to do what we call a facilitation study, which is a short-term, you want to test your method first — which is what we’re doing now, to make sure that in a one session treatment protocol, you see if there’s an immediate effect, even if it doesn’t last for very long. That’s to me, is a very good way, at least in my area of study with adult aphasia patients, people, to connect theory with the treatment and in the middle of that is the evidence you’ve established. In general, I feel that our field does call for evidence-based practice, which is very good, but I’m also very much an advocate for having a theoretical base for that, as the field is, but I think it could be emphasized a little bit more.