Chapters Transcript Video Dose in Stroke Rehabilitation: The Value of Explicit Hypothesis Testing Catherine Lang, P.T., Ph.D. presents at the Johns Hopkins Department of PM&R’s Grand Rounds on October 19, 2017. So uh so the good news is we are just a packed house. So thank you all for coming and making time today. This is our seventh annual Research and Clinical Expo with the Department of Physical Medicine and rehab. And it is by far our largest, there are 75 posters representing throughout the health system, faculty staff and also some of our academic partners that we work with as well throughout the region. And so I want to thank all those who presented a poster today and many of those posters are posters that have been shared at other national and international forums. So, thank you all of you who presented today. Thank you. The second group I wanted to recognize is Lauren Miller, where's Lauren? She in the room, Lauren's way back there in the corner, but Lauren has not been doing this for four years. She is, she makes this all happen. She takes in the posters with the committee, a phenomenal committee of People organizes it, everything that you see from poster printing to the clips that are holding them up to really vetting and working with all the people. Lauren is just a huge effort on that. So I want to just thank Lauren and the committee that and then there are a lot of unsung heroes, our rehabilitation technicians and our support staff within PM and R that help just put everything together over the last couple of days. And then someone does need to take all this stuff out of here. So people don't get upset with us and they are a part of that as well. So I just want to thank that team for just stepping forward as well. So, uh with that said, I want to ask doctor selk to uh come up, say a few words and introduce our uh grand round speaker for today. Wow, what do you think? I mean, this is uh very impressive, isn't it? I, I listen, um I'm very excited that all of you made it to come here. I'm sorry for those who had to stay behind taking care of patients. Uh Maybe next year, we, we change that. Um But this is uh very exciting. This is a very special occasion where we have uh the opportunity to show um the work that we are doing, the innovative work that we're doing in the clinical LQ I and research space. So very exciting. Uh that is such a large display of work. And I really appreciate that uh you guys push and put those posters. I really appreciate that you to get rid of time or the busy day to come over um to present and display what we are doing in the department. And I think this is a, a fantastic demonstration that the department is uh moving along. So I'm also uh want to introduce our speaker today is part of the Grand Round Series, whichever is invited to continue coming to the ground run series. Um So just a little housekeeping thing, those who are claiming CMES need to text uh the number as you know, and, and the activity code and that's how we get our CMES in place. And with that it is, I switch to that hopefully. Oh man, OK, hold on. Maybe I'll give you a little more, few more seconds to uh text something. So uh um so let me introduce uh Catherine Lang. I'm very excited uh that Catherine uh came over here. She's a, a good friend uh collaborator. I met her through the years of research. Um And she has a, a very uh nice impressive career and a very interesting work that she has been doing. So I thought we thought this would be a perfect speaker for us to, to, to bring uh for this special occasion. So, Doctor Lang is a, is a professor in uh in the program of physical therapy, occupational therapy in the Department of Neurology at was, and um she did her PT uh training in Vermont and then she did her phd at uh was university. And the interesting story here is that, uh, her advisor for the phd was Amy Bastian who is part, uh, also in the department. And I understand that Catherine was the first phd student, know Catherine. So of Amy. So it's a special relationship. Um, there. Yeah. Right. I didn't say the year it was, uh, 1982. Is that right? That baby. So, um, and then she did some, uh, post doctoral training but eventually she threatened to you and, and uh I didn't mean to minimize the post, but she, she did uh she threatened to wash you and she remains there and she has been doing a very interesting work, I think in the area of understanding those of rehabilitation, uh and in, in particular, in, in patients with stroke, uh and um in particular motor impairment and I think she's gonna be talking quite a bit about that. But also other areas that are very interesting in her work is the relationship between uh impairment, how much we change, how many impairment and how, what the impact on, on fun and we have. And uh besides other things that of course, are, have been happening in the lab. So it's very busy la very important work and I'm very glad to have you uh that I'm glad that you decided to come and join us today. So, thank you. So, I'm honored to be here. Um Coming to Hopkins is a little bit like um visiting your, your cousins and your friends because, um, you have, um, lots of nice people here who you've known for a long time and it's fun to be here. I'm also completely flattered that the room is full. So I'm gonna take that personally, whether I should or not is irrelevant. Um, and it's going to make up for the other places that I go where I have, um, just a few stragglers in the room. Um, I'm not gonna sell you anything today. Uh I can get you this. So um I am a product of, of those people that came before me and you'll see Amy's picture up here. This is one of my favorite pictures of you. Um We had a really good time when I was in the lab and she's the one that um helped me get started. Um Some of you will recognize some of the other faces. So these are the people that I've trained with. Oh I keep pushing that exciting. These are the people that I've trained so far and you'll recognize Kendra Cherry who's a postdoc here. Now, these are people at my institution who I collaborate with and then lots of other people around the country that I also interact with. So I get to be the one to talk about this data today. But keep in mind that everything that I'm talking about is the work of many people. Um I'm from the program in Physical Therapy at Washington University and all of us in our department study the human movement system, which is a system of systems that support um movement and the um field that I'm interested in is stroke recovery and rehabilitation, which is essentially a problem with the cardiovascular system and what its effects are on the nervous system. I have um three things for continuing education that I'd like you to take home today. Um We're gonna talk a little bit about active and inactive ingredients and therapy interventions. We're gonna talk about how you might apply the information um from our recent clinical trial and other studies um to dosing and rehabilitation. And then, um, at the end, we'd like to differentiate between um capacity or what people can do versus performance, what they actually do. So we're going to start with some initial observations and associations and for those of you that see different providers for your own health care in the room, you clearly don't want any of these other people to ignore the dose of medicine that they give you. So maybe they don't know how much antibiotics to give you. So let's give you a handful of pills or let's give you a few pills or you're trying to manage your Parkinson's disease. I don't know how much leva dope you should take, let's just take a handful of drugs. And so we kind of do that unfortunately, in rehabilitation and I'd like to make the point. I don't think it's going to be hard to convince you all that with a lack of understanding of dose, we have waste. So if we are um giving too much of something, we're wasting both the patient and the provider's time and effort. And if we're giving too little, we're also wasting the patient's provider um time and effort. The challenge as we move towards the research arena is that we might be wasting a really great potential opportunity. Maybe you've discovered the best intervention in the world. But if you're not giving it the appropriate dose, then we're going to waste everyone's time, energy and money and it's going to fail because it was the wrong dose versus the wrong intervention. And finally, we know that in rehab, we're making people a little bit better and we're hoping that somewhere down the road there comes some probably called a silver bullet. Earlier, I'm thinking about a magic pill or injection and when we get that thing, we're gonna need that and we're going to need to pair that with habilitation to help people become functional. And so um this line of work that I'm talking about really started with the idea of if we could figure out what dose would help people get the most better, then as new interventions come online, we'll know what to pair those new interventions with. And so let's think about this from a pharmacological perspective, you take a drug, um This happens to be Ibuprofen, every drug that you take has an active ingredient. Every drug before it's tested in humans, we understand its mechanism of action and it, how it acts on a specific therapeutic target and then it um produces the desired outcome. In this case, Ibuprofen gets rid of your headache or your back pain or your knee pain. Every drug also has inactive ingredients and you have to package the active ingredient with the inactive ingredient in order for the patient to swallow it. And I'd like to make the argument that this, we do the exact same thing in rehabilitation. So there may be some parts of your intervention that are considered your active ingredient. And then there's other parts of you to mention that your, the your patient is never going to participate in the therapy or benefit from the therapy unless you're delivering those inactive ingredients in my lab. Sometimes those come with a cheery therapist who's friendly, who asks how you're doing. You know, there's a lot of social interaction that's probably a uh an inactive ingredient and there's many other ones as well when we think about applying the different parameters of dosing to rehabilitation. Um This is a table that we created a few years ago. On the right hand column of the table on your, your left, my right hand column of the table are the list of things like active ingredients, mechanisms of action, desired outcomes, pathways, etcetera. And on in the middle is the type of things that we have to know in order for a drug to be approved by the FDA. And on the far side is some of the things that we don't know in rehabilitation. So we often don't know the active ingredients of our intervention. We very rarely know the um mechanisms of action. We do know the desired outcome. We don't know what the half life is of what we're giving. Um We don't know the pathway through which our active ingredient increases the desired outcome. Um Side effects, we don't have to typically worry so much about in rehabilitation as we do with drugs or toxicity. Um And we have no idea how the intervention that we're applying interacts with many of the drugs that our patients might be on. It's only when we know all of these different parameters, can we adequately provide the right dose of the intervention? So the best example of this from the rehabilitation arena is muscle hypertrophy. And so we have some uh reasonably well known active ingredient. We know the mechanism of action. We know the pathway, the half life, um the outcome, etcetera. Am I too close? Maybe this would help? Let's try that. I'm just doing it to keep you guys awake. Um And so when we decide that we want to strengthen someone and hypertrophy, the muscle, we know exactly what the parameters of dosing need to be. We know the frequency, the intensity and the timing and the type and that's what we don't know about most of the things that we do in rehabilitation. So now we're gonna take um that idea of these different parameters of dosing and we're going to flip them to thinking about people with stroke. Um And many of you in this room have probably heard multiple lectures about um plasticity, neuroplasticity and what's been derived from animal models. What we can see in people. I listed some papers up here. But some of the key things that come out of all of that literature are that the active ingredient seems to be task specific practice with sufficient challenge and engagement that the amount of active ingredient is large. We have no idea how much, but we know that it's much more than what we see in routine clinical practice. There may or may not be a critical period of um responsiveness. Certainly there is in the animal model and that's currently being tested here in the human model. And our knowledge of plasticity is what we call phenomenological. Like we know the different things that go on, but we don't know the mechanisms of action. Um the the therapeutic targets, the half life and all the pathways. OK. So now we're just gonna flip to talking a little bit about a study that we did a few years ago um where we looked at the dosing of um rehabilitation in people for people with stroke and what we observed is that there are small amounts of practice provided by the clinical services. This was by PT and OT and it was by seven different sites around the United States and Canada. There's no particular magic if you live in Saint Louis versus you live in Vancouver or Chicago or Boston. Um and it turns out that um the dose that's being delivered. So here's our upper limb task specific training. About 51% of observed sessions had any task specific training. This is sessions that were trying to rehabilitate the arm or the upper limb. And then we saw an average of 32 repetitions per session. And that's not 32 repetitions of the same task. That's 32 repetitions of any functional task with the upper limb. You can see that with git, we're a little bit better. So about 84% of the time when sessions were addressing the lower limb, they were practicing git and in this case, they were doing an average of 357 steps, which is an order of magnitude higher than the arm repetitions. But compared to the gate steps in animal models, it's still far below because those are typically in the thousands instead of in the hundreds. So overall, we are not providing much practice whether people are in inpatient settings or outpatient settings, whether they are being seen by occupational therapists or physical therapists, whether they're being seen by new therapists or old therapists or anything in between the, none of those things seem to really change what we see. We've done a couple of studies now, as have others. Looking at this kind of general relationship between the amount of training that you receive. In this case, the number of repetitions of task specific practice and how that rate changes um functional outcomes. So here we have the total repetitions of upper limb functional task practice on the on the X axis. And then the change in your arm function as measured by the action research arm test on the Y axis. And what you can see is that there's a nice modest correlation such that those that receive more generally do better. Um We have replicated this in a sample of people in inpatient rehab. And um a colleague has replicated this in a sample of people with gate training. So there seems to be a modest relationship between how much people get and how much better they get at least from an observational perspective. Um So the next thing that we did um is we started to look across a whole bunch of um studies in the field of stroke rehabilitation. And so this is um before we were looking at correlations, now we're looking across a whole bunch of studies. Um so much larger numbers of patients and we look to see whether getting more therapy resulted in better outcomes across rehabilitation trials for people with stroke. And we had to use time scheduled for therapy as our best proxy of dose. And I'm sure that I would be easily able to convince you that just because you have 45 minutes scheduled for therapy doesn't necessarily mean you get a certain amount of practice. But it's the best thing we've got because that's what's reported in the literature. Um We looked for studies that compared higher doses to lower doses regardless of the intervention and we included a variety of interventions and time post stroke. This is what the traditional meta analysis looks like. Um And there's a nice effect but it's actually nicer to look at it from this view. So here we're looking at time scheduled for therapy on the X axis and um the change from baseline to the to the end of the intervention on the Y axis and the metrics on the y axis are the effect size. So the bigger the effect size um essentially the greater the change and the dots that are in blue are the control groups and the dots that are in orange or brownish color are the experimental groups. And um the bigger the dot the larger the sample size and what you can see is a nice relationship between the amount of time scheduled for therapy and the effect size that was seen in those particular studies. Some of the call out buttons that are on there. For example, the B one B two and B three are the three groups in the leaps trial that was done by Pam Duncan and published in 2008 11. So their um control group, which was usual and customary care is right here. And then there are two experimental groups. B two and B three are people that got home physical therapy at a very high dose. And people that got um task specific um gate training with body weight supported uh support on a treadmill, plus some overground training. So you can see that there's a nice relationship here between dose and and um outcome and you can see that that it's a modest relationship. So it's not something that's predicting all of the changes that we see. But this is suggestive of the fact that those people who have bigger doses get a little better than those people who have smaller doses. Um For those of you who like data science, um you can actually explore this yourself. The website is right here. Um And if you, if you can't write that down, you can always email me and I can send it to you. Um This is managed by Keith Loy who is now at the University of Utah, um a really bright young assistant professor and he's compiled all of this into a database and you can go into the database and you can look at what different effects of time, post stroke or age or other variables that you're interested are and how they affect various effect sizes. The beauty of this is if you're thinking about clinical trials, you can see what are the effect sizes of trials that have come before and how much change do you really need to generate in order for it to be better than the other interventions that are already out there? So it's a really good resource for those of you who are interested in stroke clinical trials. So I've told you about some correlational or observational relationships between people who are getting more practice, getting better. We've looked across a whole bunch of studies that weren't designed to ask the question of is more better. They were designed to answer other research questions. And when you combine all of those together, um you get a general idea that getting a little bit more, gets people a little more better. So we actually ended up testing that hypothesis and um what I'm going to tell you or show you here is the value of explicit hypothesis testing and um how you can be really, really wrong. Um But wrong can sometimes be really interesting. So we set out to test in a phase two trial, we tested whether or not larger total doses result in better outcomes and smaller total doses. And then we were interested in whether or not um various factors that come along with our patients affect the dose and the outcome that they receive from those doses. So all of, you can imagine there's not going to be one magic number that every patient receives but based on their clinical presentation and the presence of other perhaps non motor motor problems or age or other factors that they might need different doses. Um We had to make some design decisions because no clinical trial is perfect. Um We because my lab has been interested in upper limb function, we chose to study that instead of um other functions. Um upper limb function is a nice thing to study when you're thinking about just the nervous system. Because if you study walking, then you got to pay a lot of attention to the cardiovascular system as well. Um We have faced some challenges and, and those questions may come up at the end that the study was designed in 2009, 10 11. Um but the results weren't out until 2016. So by the time you get to the end of your clinical trial, there's new information that you might have designed your trial better, but you didn't have that information when you were. Um at the beginning, we were really interested in the effect of dose, not the effect of the natural trajectory of recovery. So we chose to study people six months or more post stroke after that natural recovery had already occurred. Um We were also really interested in possible doses um versus achievable doses. So, um we're not going to keep people for eight hours a day, at least in my lab doing our movements. We were interested in something that would be feasible within the US health care system. And so we ended up having people come four days a week for one hour a day. Um And we chose to test dose by manipulating the amount of the number of repetitions or the amount of task specific practice. We chose not to manipulate dose by the schedule, the intensity or the frequency. And we did that because of the modified constraint induced movement therapy literature versus the traditional constraint induced movement therapy literature that suggests that the schedule doesn't seem to really matter that much. Um Our inclusion exclusion criteria are here. The big difference between this and a lot of other upper limb studies is that we took a lot more people that had nonmotor deficits. So the dirty little secret of upper limb stroke trials is often that they rule out with aphasia, they rule out people with cognitive deficits, they rule out people with neglect, they rule out people with depression and so on and so forth. So we took all those people and we measured all those outcomes and then we looked to see how those different factors modified the dose response relationship the way this study was designed. Um It looks like this. I should say, if you have questions as I'm talking, feel free to raise your hand and ask, you don't have to wait till the end. Um So people came in for a baseline assessment, they were randomized to group the four groups that we had um started with 3, 3000, 200 total reps over the course of an eight week intervention. So this would be the number of repetitions that you would have had if you had participated in routine clinical practice. And all of the practice that you got during that routine clinical practice was task specific training. So it was the sum of therapeutic exercises and task specific training. And so on that we saw in our observational study, the second dose was double that. And so in the second group, they had 200 repetitions per session. The third group was triple that or 300 repetitions per session. And the fourth group, we called our individualized maximum group. And this group had the 300 repetitions per session, but they continued beyond the eight weeks until they reached a performance plateau with the idea that perhaps the first three groups that we had chosen would be simply not enough. And that this group would let us then explore further how, how much um dosing that we could possibly give and how beneficial that could be. Everybody got task specific upper limb training, which I'll explain on the next slide. Um The um they were seen, as I mentioned for one hour sessions, four days a week or more. And then we measured each week we took some brief assessments so that we could get the time course of change. Um When, when we measure things just pre I post, we don't know if the time that we've selected is the right amount of time. But if we measure weekly, we can see when um different people start to plateau. People had an assessment, um post intervention. And then um at two months later, there's a question. So these were people that were six months or more post stroke. So they had a variety of experiences. Most of them had been through some sort of inpatient rehab. And if they hadn't had inpatient, they'd had outpatient. But the effects of that they would have finished that early in my system. The median length of stay in acute care is about three days, the length of stay, um about 30% of people get to go to inpatient rehab and the length of stay there is about 16 or 17 days. So most people, even if they go to inpatient rehab are home by the end of the first month. Is that pretty consistent to hear? Yeah. Um And then um they get whatever outpatient in patient services typically that their insurance company will provide or if they don't have insurance are provided for free by various facilities, which is um as you know, typically not very much. Um So the the intervention was individualized, it was graded and it was progressive over time. Um So each person um was given the Canadian occupational performance measure. I have occupational therapists in the room that are familiar with that. Yeah. Um And so we had them identify the things that they wanted to get better at on and the tasks that were meaningful to them. And then we set up the task that they practice in therapy to match the tasks that were interesting to them. Um The only difference between the groups was the number of repetitions that they received. Everyone had the same rules for when you grade up or down a task. A repetition was considered one activity of reaching out, grasping, manipulating and releasing things. And our goal was not to necessarily change impairment. Our goal was to the functional capacity of the upper limb. That is how do we get this person to be able to do all the things you need to be able to do with your upper extremity throughout the day. And so we couldn't possibly practice every task that they want to do. We practiced the key ones and then we hope that those translated to other tasks that they needed to do during the day. Mhm Or do you have an intuition about your your lower dose group? Maybe being able to practice the quality of the movement and you know, maybe they were trying to do the movement or maybe it afforded them the opportunity to try to do the better than you just gotta do a whole bunch. So we didn't, um we did not focus on quality of movement. We were interested in whether they could do the task for the people that we thought could improve quality of movement. We structured the task so that they didn't have a choice but to make it better quality. So a grade of is um one of our favorite task was reaching to put things up on shelves. And so for the people that we thought um could do this without circumduct their arm, we would put them against a wall. So they were forced to reach straight and couldn't circumduct. And so we manipulated, we didn't give them any cues about it, but we manipulated the environment to elicit the behavior that we wanted. And for the therapist that provided the training, that was actually the hard part, it was, it was tricky to be creative in your tasks. Um It was tricky to think about how to grade the tasks because does this person need the task graded up because they need to practice picking up bigger forces or does this person need the task graded up because they can't open their fingers wide enough? So we're not gonna do heavy forces, but we're going to do bigger objects. So we look for more finger extension. So we were using the tasks to guide the issues that the therapist thought were the impairments that were driving the function. But we weren't specifically trying to change the impairments and we were trying to structure it so that we could elicit the movement that we wanted. Um in many cases, we had people that were lower functioning. And so we were actually training their hand as the helper hand and, and in those situations, we weren't at all worried about quality of movement. Um Our primary study um outcome was the action research harm test. It's a scale that goes from 0 to 66 56 sorry, 57. I do know the scale. Um And it's a, it's a series of um items, 19 different items that test how well you can reach grasp and manipulate things. The easiest item on the test is to bring your hand to your mouth. And the hardest item on the test is picking up a ball bearing between your thumb and your finger and putting it up on a shelf. Um We had a host of other secondary outcomes and then um on the far side is the host of modifiers that we were interested in looking at. The analytic approach that we used was a little bit different than a typical prepo and Nova, it was a longitudinal analysis that they use in psychology studies. And essentially you can think about it as testing for differences in slope. So we our our groups were randomized. So they all started at the same point. And what we were interested in seeing is was the slope for the people that got the highest dose greater than the slope for the people that got the lower doses. Um And we were interested in also seeing if those slopes were affected by the different modifiers. Um This is our sample. We had about 20 people per group. Um They were well matched for age, gender and race and ethnicity. The one difference that came out between the groups is down here in Concordance. So, Concordance means that your dominant hand is the affected hand or your pre stroke, dominant hand is the affected hand. And we had one group. Um the second dose group that had a lower number of people with their dominant hand affected or a lower degree of concordance. But everything else was about the same. These are the um different baseline scores of the modifiers. Um And what we've got is kind of moderate um hemiparesis, mild to moderate hemi Preis, like most upper limb studies. We've got a range of the numbers in the print. Here are the men and the mac scores. We've got a range of scores for people that have various levels of hemi spatial neglect, depression, aphasia. This is a cognitive screener, um somatic sensory loss and um uh tone. So now what I'm gonna show you is the really disappointing results um that we were really wrong and I learned um that you always show all the data. So I'm gonna show you all the data. Um This is the lowest dose group. What you're looking at is time on the um X axis. And you're looking at the action research arm test score, our measure of functional capacity on the Y axis. Each of these are called spaghetti plots. You can see why that is each of these lines is an individual subject. And then the big blue line in the middle is the mean for that particular group. So there's the low dose group, there's the second dose, there's the third dose and here's the fourth dose. OK? And then if we put all those colored lines together, you can see that they all lie right on top of each other. Um When we look at estimates of slope, what we get are um numbers that are less than one. So these are changes per week on this test because these assessments were done every week. So what you're getting is um the max slope that we had was a little over a half a point a week in the, in the um individualized maximum group. And the lowest one we had was in the 6400 group, which is right at basically zero. So these changes are not large and um there was absolutely no difference in the slopes between groups. There were some differences in the intercepts. So this analysis also models the starting point and this should come as no surprise to the clinicians in the room. People that started with a little bit better baseline score um or people that were a little better started with a little better score that we, because we actually modeled from the second time point, not the first one, people that had more tone started with a lower score and people that were affected on their dominant side actually started with a higher score. Um But those modifiers didn't actually affect the slopes. So there were no time by um modifier interactions and there was no, there was one time by group by modifier interaction. Um This is called a what I would call a spurious finding because it makes zero sense clinically. Did somebody have a question? OK. I'm a little concerned and we can talk about it today, but you just said that one of your groups had no slow parade. It's a zero. And then uh I should not have said that there is actually a significant difference between slopes. Um So right here. So this is a great example of statistical difference, but probably clinically meaningless. So, um there is a statistically different group uh difference between this dose group, the 6400 group and the largest dose group and the smallest dose group which are higher than the second dose group. But I mean, what is what it is if you try and test for a dose effect, you want your treatment to have an impressive effect on the actual outcome measure you've chosen? That's right. And basically how can you speak to dose when the effect of any dose is essentially authentic on the A. Yes. Yes, yes. And we thought, yes, I see exactly what you're saying. Um And this is why we test the hypothesis because we thought that we would see a small dose effect on the blue line or the lowest dose. And then we thought we would see a nice dose response relationship such that the bigger dose that we got, the bigger effect that we saw, we did not see that, but it doesn't speak to dose, it speaks to the treatment. It doesn't really work on the a rat potentially hold that thought because it gets worse. Um Our our only significant effect in terms of um group by time by modifier was that the second dose group um had an interaction with depression such that those people were that were more depressed, had a slightly um steeper slope, but the slightly steeper strope was 7/100 of a point steeper. So, um again, clinically, absolutely irrelevant. Um Most people or many studies when they look at um these things instead of looking at a slope or amount of change over time, they look at the gain from pre to post. So I'm showing you that data right here. Um And what you can see is that there's no real difference between groups, everyone gained an average of more than five or six points, which is um in the literature, considered a clinically meaningful difference. And we're going to debate that in a minute. So don't raise your hand yet, John. Um, and what you can see in those parentheses is that each group contained people who didn't change at all. So the, the white values in the parentheses are the men in the max change scores that we saw every group contained a few people that didn't change at all and a few people that changed a whole lot in every trial that we do, we look at whether or not the change that people make is meaningful to them. So they answer a Likert scale that's centered on zero, meaning they don't think that their arm and hand function is changed. Plus one means that they feel like it's a little better, but it's not meaningful to them. Plus two is, it's a little better and it's meaningful and plus three is a lot better and then there's, of course, the negative ends of the scale. And so, um, what you're looking at here is the average score for the people that, um, picked each of those categories. The, the interesting thing is, is that 90% of people thought that they were a little better and it was meaningful or that they were a lot better, but they had the same magnitude of change as the person that thought that they were a little bit worse and it wasn't meaningful to them. So there's no way to get a MC ID value from this data because the anchor doesn't distinguish between people who thought they were a little better and people who didn't, and what I think that we're seeing in this picture is the benefit of the inactive ingredients of our therapy. So our folks were six months post trope. They came in and they saw a lovely coordinator. They saw really nice therapists. I was like the third backup therapist and then their behaviors went down when I did the treatment. Um But um it's really interesting that getting people out of their homes and keeping them active makes them think that they're better. But we haven't actually seen a relationship between their perception that they're better and their ability to use their arm for functional capacity. So this gets to um John's point, interestingly, at the same time that we're doing this, um we are also collecting data from our outpatient therapy services. So in that case, these are not people that are, these are all comers, these are people that um receive usual care in outpatient therapy compared to our research population, which is very carefully selected. We've spent countless hours and dollars making sure that our intervention is delivered with good fidelity. Um And on average, um if you take those numbers that I showed you in the past slide, we have a slope of about 0.5 points per week, which if you multiply that by 4.3 weeks in a month, you get an average change of about 2.1 points per month. And in our outpatient sample, we see about an average change of two points per month. So we aren't doing anything special in our research intervention compared to what's happening in routine clinical care. Um I'm gonna take a diversion for a second here because I want to talk about how we measure progress with our patients and some of the assumptions that we make everyone um that has any therapy education is familiar that knows that this is the IC F model. Um We know that we can measure body structure and function. We know that we can measure activity and what we're going to talk about is distinguishing between the capacity to do things which we measure with the action research arm test and what the World Health Organization calls the performance of that. And we can measure performance in two ways. We can measure either perceived performance, which you might capture with items on the stroke impact scale or the motor activity log. And we can measure um we can measure a more objectively performance um with body worn sensors. And that's what I'm going to talk about right now. So we spent a number of years developing ways to do this. Um For those of you who aren't sure why we would want to do this. Um Most people get referred to your therapy services because they actually want to be better in everyday life. They don't really care if they can do something better in your lab. They want to be better at home and do the things that they want to be able to do. Um, our assumption that the standardized assessments that we do in the clinic carry over to gains in performance in real life. Um There's actually a little bit of emerging data that supports that. That's actually not true. And I'm gonna show you some more that suggests it's not true. Um And then as you all know, all our self report measures are subject to bias. Um So we have been using accelerometers um to um strap on your wrist, they look like big wrist watches and we measure um people's arm movements in the hours that they're not in the clinic with us before we um collected our data in this clinical trial. We figured out that we better have some data from normal people. Otherwise we weren't gonna be sure how to interpret the data that we got from patients. And so, um Ryan Bailey who is a phd student with me, um did this work and um what he showed people were censors for 24 hours. Um And you use your dominant hand about 9.1 hours a day, you use your nondominant hand about 8.6 hours per day. So it's very similar. They're statistically different by about a half an hour, but most of the time you're using your hands together um the correlation between using your right hand and your left hand or your dominant non doin hand is, is close to one and we can make something that we call an activity ratio or a use ratio. Um that has a value of 0.95. It's simply the hours of use of your nondominant hand over the hours of use of your dominant hand. Um And this variable is tightly controlled in a healthy population. So whether you're a couch potato, um and this is what this shows here. Um So the, the number of hours that you use your limb every day is on the X axis. The use ratio is on the Y axis. And this tells us that whether you're a couch potato or you're really active most of the time, um Both hands are um moving and involved in the things that you do. I'm gonna um for the purposes of time, I'm gonna skip over this. Um And so we took the accelerometer data and I put this picture up there for those of you who know was where they are busy. Um modeling the world with functional connectivity as we were mentioning earlier and mapping the human brain. And so we've been mapping normal movement. And um what we end up with from um if we were to put sensors on most of the people in the room is something that looks like this, this tree picture. It's like a blue Christmas tree or a blue Hanukkah tree maybe works. When I talked in Chicago, I was told the blue is for the Cubs. Um I don't have orange for the Orioles here. But essentially what you're looking at on the x axis is the relative contribution of each limb to the activity. And on the y axis, you're looking at the intensity of the activity and then the color is how much time during the day you spend in that activity. So brighter colors you do more often. So most of your activities are happening are small, low intensity movements that we know from our work that actually are typically movements that are happening right in front of your body. Um But you're actually contributing, both of your hands are contributing to different activities. Whether they're not necessarily doing the same thing, you might be holding a pot and stirring with one hand, reaching up and getting something down here, reaching up and getting something down there. Any of you who have worn a sling with an arm injury, anybody had that happen to them, you know what a pain it is to go about your daily functional activities with one arm and these are just two other um healthy people. So we can look at these um in lots of different populations. We wrote a review article about this. Um Up at the top is an older adult who's in a skilled nursing facility who has a variety of medical conditions and what this flattened tree shows you is how inactive they are. It's all blue colors, but it's still symmetrical because they don't have an asymmetrical injury. Um The item in the middle is a typically developing eight year old. You'll notice for those of you who are perceptive that now the Y axis scale has changed and we had to do that because these guys are really active. And then right below that is an eight year old with hemi. And this child is actually quite active in their daily life. But you can see a slight asymmetry to their upper limb movements. We are starting to see in our data that um this issue of your functional capacity or what you can do seems to be different from what you do do. These are three participants that um from our baseline data. Um The first participant is the lowest functioning person. The second participant has moderate function and the third participant has high function. But what you can see from the first participant and the second participant is even though one has a much greater functional capacity as indexed by the action research arm test score, their actual plots of what they're doing when they're not with us, look quite the same. So it may take a lot of effort to drive someone to, to change their behavior at home. Um We know that these accelerometer measures are responsive to change. Um We have um put them on patients in inpatient rehab. This is someone that um has moderate Preis about 10 days post stroke and here's what they look like again, 30 at day 33. So um you can see that the tree shape emerges and they're able to engage their affected hand into their daily life. I should also tell you that what we see. I don't have a slide of this. But what we see in this sample is a really interesting phenomenon and that is a positive correlation between their use of their affected limb and their unaffected limb. So we had thought that you'd have a negative correlation. That is those people, people that are more impaired would use their good arm for everything. What we actually see is a positive correlation such that if you can't do much with your affected hand, you don't do much with your unaffected hand either. So it's an overall decrease in activity. And we've seen that now in a couple of different samples. So that's telling me that it's probably real um back to the trial results. What I'm going to show you is that we actually saw no change in any of our measures um that we collected from the accelerometer. We divvied up our sample into the people that we had made a six, more than a six point change and less than a six point change based on the published values of theoretically important differences in the literature and none of them had any gains on their accelerometer metrics. Um This is just an example for you. This is a 60 year old who was affected on the dominant side. This person made an 18 point change on the action research arm test. And if I had seen this person in the clinic, I would have been really proud of myself as a therapist. Look how much better I got this person. There's baseline, there's week four, here's the first post in uh assessment and there's the second one. So all of those pictures look exactly the same. I feel. Uh-huh, they certainly did. Um I'm just going to show you one of the metrics. Um This is the same picture type of picture we were looking at before where the different colors represent the different groups. The gray shaded area is um plus or minus two standard deviations of the reference range. And you can see that the groups did not approach that reference range and those slopes are relatively flat. Um If we look at six different metrics because we didn't want to put this out with the wrong metrics. So we collected a whole bunch of them. Um And we look at all of them and no matter which one you look at, you get basically the same result and that is that these slopes are essentially zero. So, no, so even though some people have changed their performance in the clinic, no one has changed their performance at home. So I'm starting to run out of time. So I'm going to give you a summary of the trial. The trial was conducted as planned. I think it was actually a really good test of our hypotheses. Our hypotheses were completely wrong. Um Despite the correlational data and the meta analysis data, when we explicitly test the hypotheses more was not better. Um Our improvements in capacity were incredibly small. We saw no consistent dose response effect and all the variables that we tested did not modify the dose response relationship. Um As I mentioned before, 90% of the participants felt that they had made a meaningful change. But those perceptions weren't related to the actual change that they made, nobody changed their performance at home, which leads us to a conclusion of intensive progressive task specific training for upper limb persis is not really worth the effort at six or more months, post stroke because we're not getting a change that's beneficial to them. You can imagine this doesn't go over well with most audiences. So I'm, I'm OK with that. Um So I'll stop here and I'll take your questions. So I think while you may not have seen the changes you wanted to see. This is a really great um behavioral activation intervention um among the group of psychologists. OK. So what were your changes in depression scores? It sounds like people were perceiving higher confidence, they were receiving good changes in their performance their mood was probably improved. We actually took those pre and post and they weren't, they weren't any different. Really? Yeah. Sure. What if they're doing the same thing? So, they're not doing more, right. They're not doing more at home, correct. So, the counts of, of bilateral and unilateral. Yeah. But what if, what they're doing is less, it easier for them, less costly in terms of energy. Whatever, what if, what if there's a change that this metric doesn't pick up? I think that, I think that's absolutely possible. I, you know, we, um, oh, so the question was, um, maybe, maybe our accelerometer metrics are not picking up changes that were actually occurring at home and I am not ruling that out as a possibility. These are, um, not perfect measures. I think they're pretty good. They, yeah, they, they activity counts of accelerations. Um, I was hoping that we'd see them start that we would see something budge. So I showed you six metrics, but we had 30 other ones. They didn't budge either. Um, these, you know, it's kind of like you can measure lots of things, but you're not sure whether what you're measuring has any meaning. So we, I'm only showing you six because I'm not sure the others mean anything. Um, but we have played around with, um, doing small things in the lab to see how much it takes to change the metrics and it doesn't take that much. These people don't move that much. So, even if you move just a little bit more, then you should see this budge and you should see it budge consistently. So occasionally we had people, um, which you can, you don't see in the, um, group data, but occasionally we had people that would be a little better, a little more active one week, but it was never sustained over time. Unfortunately, this is related to maybe three. So that particular patient show dramatic change in or to what? Really? Yeah. Yes. So the other thing that we did with these participants is um we had um an effort, a concerted effort for every participant in order to help them incorporate what they were doing in therapy into what they were doing at home. We had to be careful that we, we asked them not to practice the tasks that they were doing in therapy at home because that would have messed up the specific doses that we were delivering. But we did have a formal, a formal intervention and an informal intervention. And the formal invention intervention was that we talked to them on a specific schedule about what were they doing at home? Were they incorporating their arm more into their activities at home? Were things at home getting easier? We help them problem solve. So for example, if someone was working on, wanted to work on handwriting, which a fair number of our people did, we didn't let them practice handwriting per se. But we suggested that they try and use that hand to make out their grocery list for the week or other things that would naturally occur as part of their day. So we did do that that which is probably a little more than happens in routine practice, but probably a little less than the explicit transfer package that you see in the CIMT therapies that did not seem to translate over. Um Lots of people tell you things because they want to please you. So they say, oh yes, of course. I'm using my hand more at home where you really are or not is a completely different issue. But there's a very strong bias both in rehabilitation and any other field for people to provide, to give the providers the answer that they think the providers want to hear. And that's kind of one of the inactive ingredients in, in this and why we have a challenge with um some of the self report measures. Yes. So the the treatment, the training which may not capture the and and in the number of repetitions you say. So it may be capturing the quality of the movements or the, the endurance and then it has a specific training. So if somebody improving the Yes, yes. So, so the so yes. So this is, you know, the challenge of assessment in any rehabilitation is the items on your assessment are considered candid items that represent somebody's ability to reach grasp and release. Um We also collected the their own self report scores on the Canadian occupational performance measure in terms of how well they thought they perform the task and how satisfied they were with performing the task and surprise, surprise. Everyone thought that they were better at doing things and everyone was satisfied, but there was zero dosing effect in that. Um And if you look at lots of other studies with the Canadian occupational performance measure, everyone always gets better on that measure because they've come in and invested time and energy. So, so it's possible that there were people as Amy suggested that we're moving with a little bit better quality or little less effort on a particular functional task at home that we weren't capturing, but it's probably not going to be enough to make a big difference and not enough to give you a metric that your insurance company is unfortunately going to pay for your services to your friends. It's a very interesting study. There was actually a recently published article um complete final. Also looking at a similar aspect of like looking at this competitions with like step. Yeah. Um like a local mac group just walking into the ground and they found their hypothesis was that they had the patients on the would be better because it seems like they were getting more steps. It was actually through walking over ground better, but they had less steps. So it was this like component of a less dosage but do you think there's this component of like intensity that were missing? Like, so in a year, do you, do you remember who the authors of that study that you were talking about are? Oh, ok. OK. I was thinking it was your groups. Um So, so yes, so, so I think that there's, I think there's a really big difference between walking training and upper limb training because you don't need that much strength in your muscles to lift your arm and you don't exert that much cardiovascular resources to move your arm compared to having to control your body weight, exert forces to control your body weight and have the cardiovascular conditioning to walk. So I think that that there's different components when they walk on the loco mat, there's some nice data from Chicago, from the lab that Kristen Lee used to be in George Hornby's group that suggests that being on those devices, you actually reduce your effort because the machine does it for you. And so that, that may be exactly what we're seeing in that um in that study is that the training, actually, you use less effort in the training and therefore you got less benefit. I'm not sure that those same intensity issues translate to the upper limb, but it's certainly a testable hypothesis that would be really interesting to figure out. So the individual tasks were not harder. And so that's what we took great pains to, to do is we had a set of rules that the therapist had to follow related to how difficult each task was in terms of motor control difficulty. Um It's possible that the people that were did more repetitions felt that it was more intense because it was longer but not the individual repetitions of the task that was very carefully controlled because we could have gotten ourselves in a situation where the people that are in the lowest dose group actually ended up with the hardest tasks, they had more time to do them. And the people that were in the highest dose groups got the easiest ones, but we worked really hard to calibrate so that we were working at a specific challenge level and that, that was consistent across groups. So there was another question in the back. OK. How similar was your intervention to the other interventions that you would look at from that analysis? Like, is it possible that there is full dose effect if you didn't do the right intervention? Anything is possible. First of all, um anything is possible, there's some really interesting data um that comes from three separate countries. So are you familiar with the proportional recovery rule? OK. So there's a, there's a rule that was initiated in Colombia when Dr Krakow was there. Um It's now been replicated in the Netherlands and it's been replicated in New Zealand for the upper limb that suggests that people recover about 70% of the available range that they could recover. And you can see this in America where our rehab system is ok. You can see this in the Netherlands where their rehab services are Cadillac services. These guys get much more than we do. And you can see the same thing in New Zealand, which probably falls somewhere in the middle. And the fact that you see the same degree of recovery of impairment across three different countries with three really different health care systems suggests to me that what we're doing is as therapists is we are capitalizing on the impairment gains that are coming through natural recovery and helping people incorporate that into function. But we're probably not changing that fundamental level. Um I think that's exactly what we did in this dose study. So we had people that were six months or more post stroke and I think the small gains that we saw were because these people had some latent abilities that they were capable of using but weren't actually using. And then by being with us, they used them a little more when they were with us, but they didn't use them when they weren't with us. That's, that's my kind of biased reflection of what this data is telling me. What do you say are the next step, the next step? Great question. So, um one of the things, what just one last question. OK. So one of the things that we're really interested in is this discrepancy between capacity and performance. And we've actually collected another outpatient sample um up in Chicago. And um what we see is we see kind of three patterns in um outpatient, the services. The pattern over here are people who you can't see the numbers because they're too small. But the first person had nice change in their capacity to do things and they also improved their performance at home. So you can see that shaped tree becoming more normalized from the top down to the bottom, the middle group of people are people that um change their capacity but showed no change in performance. And then unfortunately, we saw a number of people that had no changes in capacity and no changes in performance. And what we're really interested in is exploring this issue a little bit further. Um So it's a huge assumption that we teach our students um at least in physical therapy and occupational therapy, I can't speak, speak for speak language pathology, but we certainly teach our students when they cover um and learn how to manage patients with all different diagnoses that the changes that you see in the clinic that we make you do with these standardized assessments suggest that or imply that people are getting better at home. And I think that we have to examine this issue really carefully. I'm sure there's other fields that have already done this better than we have, which is typically what happens in rehab. So we're interested in knowing whether this is a phenomenon that's unique to the upper limb. Or do we also see it in people with mobility problems, post stroke, we're interested in learning whether this is a stroke problem or it's a general neuro rehabilitation problem. And our candidate other population is people with Parkinson's disease. So people that get referred to therapy services with the expectation that the therapy services are going to help them be more mobile. And we're going to see if that's the case. And then we're also interested in the trajectory of these trees over time. So we all have seen tons of data with those impairment curves in the functional recovery curves of how the most recover occurs in the first month and then starts to slow down by about three months. People pretty stable by about 3 to 6 months. We're thinking that some of the challenges with our intervention is we were happening six months, post stroke when their behaviors are already formed. So they've already been reinforced to do things with, to either not do something or do things with their unaffected hand. And so there may be a sweet spot where we could get in at the right time. They're also may be low hanging fruit in this middle group. People that have the that can change their capacity but may need a health behavior intervention on top of the motor intervention in order to get them to incorporate that in every day. So we're exploring some of those things now, including surveys of people's self efficacy, motivation, engagement, um barriers to using their lamb and so on and so forth. I know nothing about that stuff. So I have some really great collaborators that are helping me. Excellent. Thank you. So I, I stop this in the interest of people who need to leave. Um We have one more hour this uh for those who do need to go ahead and get some lunch, there's more posters and Catherine will be around to entertain more questions. Great. Thank you everybody for joining us today. Created by