 Functional Status and Disability in the Canadian Modestinal Study on Aging. Welcome to our speaker, Alexander Mayhew. Alexander Mayhew is a PhD candidate in the Department of Health Research Methods, Evidence and Impact, ATI, at McMaster University. She previously completed her Bachelor's of Applied Science at the University of Guelph in Applied Human Nutrition and a Master's in Health Research Methodology at McMaster University. Her current area of research includes estimating dietary intake and large cohort studies, assessing methods of defining muscle mass loss in older adults and understanding pattern statistical function and disability limitations. Again, before we start, let's remember that there'll be a question and a session at the very end of the webinar, but please feel free to write at any questions or comments at any time during the webinar and the chat box, and we'll get to it at the end of the session. Just as a reminder, when I pass over to Alex to start her talk here, she will be speaking, even though it's under Catherine Galley's hosting capability. So I will pass it over to Alex Mayhew to begin her talk. Hey, thank you to everybody for joining us today and to Carol for the great introduction. As Sasha mentioned, I'm gonna be talking about functional status and disability in the Canadian Longitudinal Study on Aging. And I'm gonna be talking about how we've chosen to define functional status and disability for a project that we've been working on, the prevalence of those variables, as well as the relationship between function and disability. So the rationale for this project was pretty straightforward. We know that maintaining independence throughout the aging process is a key concern for older adults as well as health and social services. Age-related disabilities have numerous implications, such as the increased demand for healthcare, a reduced quality of life, an increased cost of care, and higher mortality. So there's been a big emphasis on trying to find ways of keeping people from becoming disabled as long as possible and having them live independently. And one of the areas of research has been in functional impairments. And that's because functional impairments are thought to precede disability and therefore may allow for early interventions to prevent or delay disability. So ideally we wanna see more older adults living as the woman on the picture on the right is, who's perhaps in her own home, she's still preparing food for herself despite being an older age. So we had three primary goals with this project. The first was to estimate the prevalence of functional limitations in the CLS, say. And we did this both in the comprehensive cohort as well as the tracking cohort, and they are measured differently between the two cohorts. So we have a questionnaire-based self-assessment in the tracking cohort and performance-based testing in the comprehensive cohort. We wanna testmate the prevalence of disability in the CLS, say, and explore the relationship between the measures of functional limitations and disability. And though there are many different models of disability available, we decided to focus on Navi's Disablement Model. It fit in best with the overall goals of our project and it is one of the most commonly used models as well. And according to Navi's Disablement Model, people pass through four different stages. Originally, there's some sort of pathology or pathophysiology problem that happens, and this could be due to disease or illness. And that may progress on to an impairment which then could become a functional limitation, which is one of the focuses of today's talk, and some functional limitations then progress on to disability. Focusing on those last two boxes, the big distinction between a functional limitation and disability, according to Navi's Disablement Model, is that disabilities stop people from being able to do the things that mean something to them in life. And you'll see on the figure here that they do talk about social aspects. So it could be that people aren't able to get out and about the way that they want to to participate socially. It could be that they're unable to care for themselves the way that they want to. And this model does suggest that there's an interaction between the environment and the individual. So some environments are certainly more conducive to allowing people with functional limitations to keep on doing those tasks that are important to them, while some environments are less conducive to that. So within the CLSA, disability is defined the same way in both the tracking cohort and the comprehensive cohort. And there's multiple variables that get at that kind of social participation element. But what we've chosen to focus on are activities of daily living. And those could be further subdivided into basic activities of daily living, which are mostly self-care tasks, including being able to feed yourself, bathe yourself, make it to the toilet in time, get dressed, move about in short distances, and get in and out of bed, as well as look after appearances. Then there are also instrumental activities of daily living, and these are more complex tasks. They don't necessarily relate to self-care, but they are necessary for people to continue to live independently. So it's the ability to do things like grocery shopping or go out and buy clothing, manage your own money, take medications, use the telephone, go to places outside of walking distance, prepare your meals, and do housework. So as I alluded to, there are different functional status measures in the CLSA. So unfortunately, this makes this presentation a little bit difficult to follow along with at times. So I'll always try and make sure that it's clear which measures we're referring to. So we had those objective interviewer-administered performance tests in the comprehensive cohort of merely 30, or just over 30,000 participants, and the subjective self-reported questionnaire in the telephone cohort, or the tracking cohort of 20,000. And we know that self-reported measures of function and performance testing are strongly correlated when they're actually measuring the same sort of underlying domains. On the flip side, performance testing and disability are only moderately correlated, but this is what you would expect because they're supposed to be getting at slightly different constructs that are related, but shouldn't be identical. It's quite interesting that there is a bit of opacity of evidence about the association of self-reported measures of physical function and disability. And this is because there's a lot of messiness in the literature about what terminology to use to describe these constructs. So when you're looking at self-reported measures of physical function, in many cases, people are actually using activities of daily living to identify this. So there aren't as many studies comparing questionnaires to questionnaires as there are the performance measures to activities of daily living. But regardless of how you measure performance, be it through the performance tests or the self-reported questionnaires, there is a strong association with disability, mortality, and other poor health outcomes. So in the tracking cohort with that physical function questionnaire, we have previously done a factor analysis to try and determine if there were any underlying domains. So we went into that without any real theory about exactly what was going to come out. So it was data-driven, but what we found is that three domains emerged. We had upper body limitations, which include a variety of tasks. Some of them require things like strength. So for example, lifting 10 pounds from the floor. And some of them are a bit more of a mobility or range of motion issue, such as washing your back. And there are six questions for that. We also found a domain for lower body limitations. And similarly, these required tasks for range of motion is important, such as scooping, kneeling, and crouching, as well as tasks that are a bit more strength or endurance-related, such as being able to stand up after sitting in a chair or walking two to three city blocks. There were only two questions for dexterity-related limitations. And that was handling small objects and using a knife to cut food. We do have the performance measures in the comprehensive cohorts. So very briefly, there are five of them. Gate speed, which is very commonly used. And in the CLSA, specifically it's a four meter walk test at people's usual walking speed. And we did calculate that into meters per second. It was a four meter course. We just divided the number of seconds it took by the meters. There's a timed up and go test where participants have to get up from a chair, walk three meters, turn around, and walk back to the chair and sit down again. A balance test. And we had participants first balance on their right leg, then on their left leg. And this was measured up to 60 seconds or whenever their other foot hit the floor or if they use the wall to stabilize themselves. And for the purposes of this project, we used the best time, regardless of if it shows the right or left leg. There's a chair rise test, which participants have to get up and down from a chair five times without using their arms. And grip strength, which was done in the dominant hand. We used the highest value of the three repetitions that each participant did. And this was all done using hand grip strength dynamometer. In terms of participant characteristics, I'm sure many of you are quite familiar at this point with the data from the CLSA. But we do have a mean age of about 63 and that's reflective of the 45 to 85 year olds that were recruited. It's about a 50-50 breakdown between males and females. The mean BMI is 27.7, so just in the overweight category. We do have a relatively educated population. So 75.5% of our participants have a post-secondary degree or diploma. And similarly, we do have relatively high income participants with a bit of a skew towards those higher income categories. And we looked at the number of chronic conditions and at least 67% of participants had at least one of these present. So first I'm gonna discuss the prevalence of disability. So for disability, we were looking at people who required assistance with at least one activity of daily living or instrumental activity of daily living tasks. And the reason we chose to operationalize it this way is that we actually don't have a particularly high prevalence of disability in the CLSA. So we felt that it was better to leave those clumped together though there's certainly some limitations that I'll discuss later about that because it gave us a sufficient sample size for the logistic regression analyses that we'll be discussing later. But what we do see is that for disability, the prevalence, there's quite a dose response across the age categories. And that females consistently do experience more disability than males do. And we found that certain tasks were more likely to be endorsed than others. So for example, getting to the bathroom in time and doing housework were two of the most problematic tasks. But it's still there's relatively low prevalence for each individual task. And this makes sense because once people have accumulated two or three or more deficits, they're probably likely to require some sort of assisted care facility. And that's in the context of the development of activities of daily living. Originally, they were used for care settings to determine resource allocation. So it was things that you're expecting people to need a lot of hands on help with. And the prevalence of the questionnaire-based functional limitations. So this is in the tracking cohorts. We've chosen to subdivide this by those domains we came up with in the factor analysis. So we have upper body tasks, lower body tasks, and then dexterity. So the slide that we're on is focusing on having difficulty with that least one upper body task. And you see the exact same trend that you did with the prevalence of disability. Overall, there's an increase in the prevalence across the age groups in both males and females. And consistently, females are having more difficulties than males are. You'll also note that the absolute amount of limitations is higher in all of these age and sex strata. So it's indicating that more people are having upper body task limitations compared to ADL or IADL disabilities. And this slide has the same setup, but it's looking at lower body tasks. So you'll notice that the prevalence of lower body task limitations is higher than the upper body tasks were. And same trend increasing across the age groups. What was interesting to see is that males actually have more limitations in the higher prevalence of limitations in the lowest age bracket compared to females. We have a few hypotheses about why this might have happened. And it could be that males are doing more demanding tasks and therefore feel like they're actually having more difficulty with them in reporting that. But that's just our hypothesis and there's no way of proving that at this point in time. And then finally, we are looking at difficulty with at least one dexterity related task. And you can note that even in that oldest age group in the females, the prevalence is still less than 15%. So this is indicating that there's not a lot of people in the sample who are having difficulty with these dexterity tasks, but there were also only two of those included in the questionnaire. So it makes sense that there would be lower prevalence than either upper body or lower body limitations for which there were six questions for each of those. And it could be that they're actually more simple tasks in terms of using a knife or picking up small objects that people are genuinely having fewer challenges with. But you see that same dose response across the age groups and then once again, females consistently are reporting more limitations compared to the males. We also looked at normative values for the functional limitations. So even though our original goal, we set out to come up with a prevalence of functional limitations for those performance tests included in the comprehensive cohort. What we found in the literature was there aren't a lot of great cut points to indicate who is actually functionally limited. And something that was really disregarded consistently throughout the literature was the implications of which age groups were included in developing those cut points. And I think our data really nicely shows that this isn't something that should be ignored because people who are in younger age groups should have higher performance than those in older age groups. And there may not just be one absolute cutoff that applies to everybody for saying that they're starting to have a problem. So here are the normative values for gate speed measured in meters per second. So in this case, you want to have a higher value that indicates a higher performance. There's not a huge difference between males and females, especially in that younger age group. However, as the age groups progress, there does seem to be a bit more of a separation between the two. And it is very clear that performance does decrease with age. And if we're looking at the literature, generally the cut points that are recommended are either 0.8 or 1.0 meters per second. And it's quite interesting to look at how that would fit in with this data. So in the youngest age group, if you were using a cutoff of 0.8, there'd be very, very few people that would be considered functionally limited whatsoever. But there would be a good deal of people in the older age group that would be functionally limited. And you'd be identifying a very different population if you use that 1.0 meters per second cutoff. So again, that's something that just hasn't really been thoroughly discussed in the literature. The second performance test is the timed up and go test. So for this one, taking more time means lower performance because it's taking you longer to get up, walk three meters and get back to the chair and sit down again. Again, you don't see a huge difference between males and females, but there is very clearly a dose response. And as people are getting older, it is taking them longer to perform this test. One of my favorites is the balance test. So this one is in seconds. So you want people to be balancing for as long as possible. And this one perhaps has the most profound decrease across the different age groups. So we find in our youngest participants, a lot of them have no problem balancing for the full 60 seconds. And the mean balance time in both males and females is over 50 seconds. But by the time you get into that 75 years in an older age group, the mean balance time is less than 20 seconds. And quite anecdotally from having done some of the data collection, there's a lot of participants who pretty much their foot comes off the ground and immediately it goes back down. So they only have a time of one to three seconds for balance. So there are still plenty of older adults who are able to balance for the full 60 seconds. So unfortunately, these graphs don't really show the heterogeneity, but there's a lot more older adults that are on either far end of the spectrum compared to the younger adults who pretty consistently all perform well. And this graph is for the chair rise test. So in this one, you want lower values opposed to higher. Very, very little difference between males and females, but you see that increasing trend as people get older. And the last graph here is for mean hand grip strength measuring kilograms, also stratified by age and sex. And this is a test, it's very well known in the literature that males and females do have different grip strength potentials with males being significantly higher than females. So cut offs are recommended specifically for each sex. This isn't something that necessarily happens for the other performance tests, but as you saw, there's not merely as radical of a difference between them. However, what still is ignored in the literature is the fact that grip strength does very clearly show a decrease across these age groups. And it may look as though the males are losing much more grip strength in females, and they are in terms of absolute number of kilograms. But if you look at the percentage of hand grip strength, or the difference in percentage hand grip strength between the age categories, it's actually quite similar between males and females. So the bigger absolute change is just reflective of the higher starting value for the males. But very clearly it decreases as participants were getting older. So that's kind of it for all the descriptive statistics. We wanted to start looking at the relationship between function and disability. And it's very clear in Nagu's model as well as other disablement models that fiscal functions supposed to precede disability. Unfortunately, we can't actually show that causal relationship. We were only using the baseline data from the CLSA. So we can't establish temporality, but there is a really strong theoretical framework for this. So for each of the models, our outcome is having at least one basic activity of daily living or instrumental activity of daily living limitation. So as I said before, we decided to operationalize it this way simply because of sample size issues where not that many people had one or the other was better to combine them. And we have four different models for both the tracking cohort or telephone cohort as well as the in-person cohort or comprehensive cohort. And these different models were trying to just assess physical function in different ways to understand if different combinations of limitations were associated with disability in different ways. And it matters if the number of domains limited were considered and what the concurrent adjustment is each of those variables were. I'll get into more detail about exactly what those mean in the next slides, but that's the gist of what we did. And all models were adjusted for age, sex, the number of chronic conditions, self-rated pain, household income, depression status, body mass index, alcohol consumption and cognitive decline. So for the telephone-only cohort or tracking cohort with that questionnaire-based assessment of physical function, these were the main results. So in model one, we just compared people who had at least one limitation regardless of which domain it was versus not having any limitations whatsoever. And we found that the odds of having an ADL or IDL disability was 3.67. And the next model, and even though it doesn't look like perhaps the most impactful model, we haven't been able to find any other literature which has concurrently looked at these different domains within the same model. And I suspect that's largely a sample size issue. For this analysis, we had still 20,000 participants to include, which gave us a nice narrow confidence intervals for this analysis, which other studies aren't necessarily able to achieve. So what we found is that each of these different domains, upper body, lower body and dexterity, are independently associated with activities of daily living disability. So this is quite an exciting finding because it does show that these domains are measuring different underlying constructs all related to function. For model number three, we looked at the number of domains with at least one limitation. And for this, we didn't care which domain was impacted, just the number of domains. And we were quite shocked to see that when participants had all three domains limited, the odds of having an ADL or IDL disability was 13.11. It is a bit of a wider confidence interval, but it's still a very, very strong effect size. And then you see that decrease down to 6.69 for the odds ratio for having any of the two domains limited and having just one domain limited, the odds ratio is 1.90. So again, this is a very exciting result for us because it's showing that as people have more domains that are impacted, you definitely see that increase in the risk of disability. And model four is the one looking at the individual combinations of the domains with at least one limitation. So not surprisingly, all three limitations, the odds ratio of 13.19 matched up quite nicely with model three with having three domains limited because those are essentially exactly the same. And then we have the different combinations of two domains impacted. And we did find that upper body and lower body limitations in combination seem to have the highest risk of disability, especially compared to lower body and dexterity limitations where the confidence intervals weren't overlapping. For upper body and dexterity limitations, for whatever reason, there's a very small number of participants in that category, hence the wide confidence interval. So you can't say if it's significantly different from either of the other two. And then certainly for each one of the individual variables, those were all very similar in overlapping with a range of 1.87 as the odds ratio to 2.71. And for the in-person cohort, so based on those performance tests for physical function, the models all were structured the same way as for the tracking cohort. So in model one, they're having at least one test in the lowest performance quintile. We found an odds ratio of 2.22. And similarly as the other model two for the tracking cohort, we adjusted for the presence or absence of the lowest quintile of performance for each of the individual tests. And again, we found that each of them were independently associated with having an activity of daily living disability. So again, a very exciting result that we haven't been able to find anywhere else for the literature. For model three, we were looking at the number of tests in the lowest performance quintile, just regarding which specific test it was. So once again, there's that very nice dose response. And in particular, it seems when people move from having three tests in the lowest quintile to four tests in the lowest quintile or five, that's where you see the most dramatic increases in the odds ratio for having an activity of daily living disability. And model four, I apologize, it's quite large. I'll also note that there are six variables that are six combinations that have been taken out due to a lack of statistical significance. And some of those were individual variables. So you'll note that the timed up and go isn't included on here. And some of them were combinations of either two, three, or four. So there wasn't a clear pattern about which results were statistically significant and which weren't. But what is shown here are all the ones that still were statistically significant. And you do see that dose response in general, that as the number of tests in the lowest quintile of performance, as they were accumulated, the odds ratio went up. But I think in general, this indicates that there's not a huge difference necessarily between the different combinations of tests. And even if we left all the data in that wasn't statistically significant, the confidence intervals were wide enough and the point estimates were high enough that they would still overlap. So that's pretty much it for what we did for the analyses. So we found that overall 9% of participants had at least one activity of daily living limitations. And this is quite similar to the estimates in the Canadian Community Health Survey, particularly when you start stratifying by age. So those age 65 and older. And it is very consistent in the literature that more females have limitations in males. Something that I'm quite curious about is, is this a reporting issue where females are just more likely to say that they're having challenges or is it that females really are experiencing more challenges in comparison to males? And for physical function limitations, we did find that they're more prevalent than activities of daily living disability. And there's two possible explanations for this. So it could be that the questions are phrased quite differently. So for the physical function questionnaire, it's very explicit in asking, do you have difficulty completing this task? Or it's for the activities of daily living. It's phrased, do you require assistance doing this task? Or are you able to do the task independently? So it's quite likely that more people are going to say that they have difficulty doing a task, but they don't require assistance. And there are a small number of studies in the literature that have confirmed this. Then there are people just saying that they require assistance, but don't find the task challenging. And this is indicative that these limitations do precede disability. So part of it could be the phrasing of the questions. And it could also be that the questions themselves are easier tasks or less complex. So people are going to have more issues with them. So you can have difficulty with stooping, kneeling and crouching, but that doesn't necessarily mean that you're going to progress onto a disability unless you've reached a certain threshold of having difficulties with that. We did find that there is a strong association between function and disability, regardless of if it is measured by questionnaire performance testing. And this was a great exciting finding for the CLSA. His unfortunately, we did not do the questionnaire based in the performance testing in any of the same individuals. So the fact that the results were so consistent between the tracking cohort and comprehensive cohort, ones itself to saying that both of these tests are probably good valid measures. And though we can't directly compare them head to head, they probably are getting close to some of those same underlying constructs. And we did find that there's an independent effective each domain of function in the questionnaire based assessments and performance based measures for disability. And again, as far as we know, this has never been done in the literature and it has important clinical implications. As it's suggesting that looking at multiple domains or doing multiple performance tests is going to be clinically relevant for helping to decide who's most at risk of having a disability. And perhaps those are the people that should be prioritized for interventions. And of course, we do have limitations. I think one of the biggest limitations is that we did only use activities of daily living to define disability. It's well known in the literature that these are limited in scope. I already said before that they were developed originally for determining resource allocation and care settings. So they're not perhaps the most relevant to community-developing older adults. And that was reflected in the generally low prevalence. And because of the low prevalence, we were unable to look at things like different combinations of activities of daily living and which performance tests perhaps were predicting which deficits in those activities of daily living. Though as our CLSA population ages and grows, hopefully we'll be able to start answering questions like that. There's also very little understanding of how performance tests and questionnaire-based measures of physical function map onto one another. And because we don't have those done in the same people in the CLSA, unfortunately, that's not something that we were able to assess. There is also overlap between the activities of daily living questionnaire and the physical function questionnaire. And this happens all the time in the literature. And unfortunately, our questionnaires are no exception. So for example, there's questions about walking in both, even though they're for different distances, it still does, perhaps those constructs are too closely related. There's also the issue of a lack of clinically validated cut points for performance testing. So I said before that there hasn't really been any discussion of the importance of the role of age. And I do think that this project really nicely illustrates that this isn't something that should be ignored and that we know that the performance tests differ. I didn't present the data for it in this presentation, but we have stratified those values by healthy participants who didn't have any sort of disability and didn't use mobility aids, versus unhealthy participants who did have disability or use mobility aids. And we found that there are big differences in values for healthy, there's unhealthy young participants who would still have a higher gate speed compared to healthy, older participants. So it seems like one threshold applied to multiple age groups just isn't going to cut it. And to really assess these, we do need longitudinal data, but I think that's something very exciting that we can do at the CLS State data in the future. And of course, we weren't able to assess the causal relationship between physical function and disability. For that, once again, we are going to need longitudinal data. So in conclusion, in this analysis of over 51,000 participants from the CLS State, the overall prevalence of disability was about 9%. And we found that functional limitations are more prevalent than disability. And within the domains, we found that lower body limitations were the most prevalent at about 41%. Upper body limitations for the second most prevalence had around 25% overall and dexterity related limitations only impacted 7% of participants. It's clear that functional status is lower than older adults, but unclear where the cut points should be and therefore we couldn't assess prevalence. And on all measures, females tended to be more limited than males with only a few small exceptions. And we did find that regardless of how you operationalize physical function, there is a strong relationship with disability assessed by activities of daily living. And I'd like to give a quick shout out to my collaborators. Without them, this project certainly never would have gotten off the ground and they've provided really important support and advice and guidance throughout this process. And now I think we'll open up the floor to some questions. Thank you, Alex. That was really a really excellent presentation and well presented. And now I'd like to open it up to questions. Just a reminder to everybody that muting remains on, but you can enter your question into the chat box in the bottom right-hand corner of the WebEx window and I'll go ahead and read it out. And we can just ask the question and procession that way. So to get started, there's a question from Jin Liu. What is the variation with age change? Are they similar of the ranges or getting larger? So I also kind of had a question about, maybe you could talk a little bit more about that heterogeneity that you discussed as age increases. Generally, there is an increase in the heterogeneity for all the performance tests with age. And I think that speaks to just the general heterogeneity and aging, which is part of the whole premise underlying why we're doing the CLSA. So as people age, you do end up having very high performers who for whatever reason have been able to maintain their performance status. And given that we only use the baseline data, we can't say if people just had much higher absolute values to start and therefore, even though they've declined, they're still much higher than average. Or if they've just experienced less decrease compared to their peers. And then you also on the flip side have people who have very low values in the older age groups. And again, we can't be sure if it's because they haven't aged as well or if they just started off at a lower value. But that is definitely observed that the spread becomes wider as people get older. Did you say that you did look at those well, those good performing older adults compared to the poorly performing older adults? We did. So the data wasn't included in the slides, but we did find, I can't recall exactly what the difference ended up being. I think in some ways, it was fairly consistent across the age groups where poor performers, the magnitude of difference between the healthy individuals and the unhealthy, the magnitude of the difference is similar for each of the age categories. So it wasn't necessarily that the poor perform or the unhealthy older adults had a huge difference in the healthy ones, but that wasn't observed in the younger people. It was fairly consistent throughout. Yeah, that'll be interesting to see on the longitudinal data. It will be. Another question from Yixing Chao. What statistical packages are used? And I'll also tack on a small question just to specify that all of your quintiles were aged and specific? Yes, thank you for highlighting that they were aged and specific quintiles. And we used data for all of our analyses, or sorry, SAS for all of our analyses. So we have time and space for more questions. So feel free to type it into the chat box and I'll read it out, but I'll go ahead and with a couple of other questions here. So how would you look at environmental factors that are related to the disability process? What measures would you consider to look at that kind of wider environmental issues? That hasn't been honestly something that I've given too, too much spot to yet in terms of these projects. I think some of the variables that you can use to measure disabilities lend themselves more so to that environment than others. So I would be very interested in looking at how these functional tests and the questionnaire end up associating with variables such as are people being held back from participating in community related events as much as they would like, given their functional status. So there's some more subjective measures that are really people's opinions about what they're able to do versus not able to do and how that impacts them, rather than just saying if they have required assistance with the task or not. So I think that would be a good starting point. And we do have some great variables in the CLSA for things like social support availability that I think would be interesting to see if they end up moderating the relationship between function and disability. Yeah, but there's a lot of interest in kind of the built environment and the social isolation pieces of all of that to be interesting to explore. And I think we need more environmental collaborators to really begin to unpack that because it wouldn't be simple or easy, but certainly worthwhile to look into. Just a reminder to everybody, if you want to write a question, make sure that you send the question to all presenters, not just the host. So to click all, all presenters are all for that question. So another question from Walter Widdich, are there specific reasons why you choose to include measures of vision and hearing in your models given their link, why you did not probably, given their link to ADLs and IADLs? That's something that we're looking at in some further analyses. So we're interested in vision and hearing as more moderating variables and that seems to be how they're coming out. So that is something that's on our radar to look into and it certainly is important. And our preliminary results, I can't cite numbers unfortunately, but the results of that analysis that we've run so far do indicate that people who have a vision impairment or specifically vision ends up really impacting disability. It was the fact that's far less so in those of hearing limitations. That would be very interesting. Mark Kay asked, did you use the CLSA weights in your descriptive and regression analysis? We did. So we use the inflation weights for all of the descriptive analyses. So the results presented today should be generalizable to the entire Canadian population. And we use the analytic weights for the regression models. So both were weighted, absolutely. A question from Lily. You mentioned previously that functional limitations include psychological and social limitations disability. How do you think these are defined and measured? So that's certainly a limitation of this project that we really were focused on the physical limitations rather than psychological or social. So I think there's a lot of thought that has to be given. It's been easier, I think, choosing the disability outcomes for the psychological and social limit or disabilities than the limitation side of it. So again, that's something that we'd really like to do, but we just haven't quite gotten the right team of people together that provide us with the expertise to be able to consider those. But it is a very worthwhile endeavor. And it does require, from what we've looked at, you have to be a bit creative with the CLSA data because there's not just one questionnaire that's necessarily going to get at those constructs for you, which is why we definitely need the additional expertise. Another question from Judy Bedel from Ying. Did you consider some factors such as ethnic groups and cultural background near now? We did. And we tried those out for the regression modeling, but unfortunately, they weren't statistically significant within the models, and that really comes down to the sample size. Our population within the CLSA is very heavily weighted towards those of European descent. So I just don't think we had the sample size to figure out anything for those other different ethnic groups, especially given that we had a relatively small number of people with the outcome of interest. But we would expect to see that there could be some ethnic differences there, and perhaps that's something that if people are accumulating more disability, or if we operationalize disability in a different way, we might have the power to start looking at that. There's another comment from Mark Kay about being able to access the PowerPoint slides. So I'll leave that for our communication director to write in the comment section about how that happens. As we move on with the questions. For the conference of cohort, did you adjust for province and your models again from you? That we did, and that's part of the, when you follow the waiting document that's available on the CLSA website, that's something that you're supposed to do. So we followed that recommendation, and that wasn't included as a covariate, so I guess technically it should have been in that list. Was it significant? Generally not. A couple of the provinces were significantly different, but for the most part, there wasn't any rhyme or reason, and we did do some pre-analyses looking at, perhaps if region mattered, and grouping the provinces by region, and that by and large is very null. So it seems it was something that was in the model for the waiting, but doesn't have any big implications or any big conclusions. We'll keep asking questions for about five more minutes here. Lori Churchmuch asked, which test results did you use to adjust for the cognitive function? So again, that list of the founders that you used. So for, I remember the acronym of that, and I'm just going to try and find if I wrote that down somewhere. So it was the MAT test, which I think is the alternating word and number test. Trying to look that up quickly. Yes, the mental alternation test, and we did a standardized score with a mean of 50 and standard deviation of 10, and participants with a score of less than 35 were considered to be cognitively impaired. Okay. Was that age and sex specific? So we talked about the provincial differences. So we have a question from Jang Lu about did you look at any geographical differences? And again, I think you included it, but didn't look at it specifically. We did see some investigative analyses. So that was the clumping of the Atlantic provinces together, the prairies together, but there's really nothing that came through. So it doesn't seem like there's some sort of pattern where certain areas of Canada are having a stronger relationship between function and disability than others. So this might go with the environmental factor question as well, but did you look at rural, rural urban environment? No, that's something that happens, and I think that's something really worthwhile to look into. Another question from Judy Beldel. Do you anticipate a time when functional exercises will become a standard of care versus rest, chronic sitting and inactivity? The research on this is getting old. Well, I personally certainly hope so. I'm always surprised that we still haven't really focused on eating well and moving as the treatment to a lot of different chronic conditions and just in general improving the life of people. But of course, it's easier to say get up and move and eat correctly than it is to get people to do that. But I do think, I would agree that the research is getting old, but I do think it's consistent with showing that really if people can do some sort of exercise, and there's a lot of great work coming out of the kinesiology department at McMaster, showing it doesn't even have to be extreme levels of exercise, it can just be walking or it can be 10 minutes bursts of exercise that can end up having huge improvements in terms of people's function. But I'm not sure, I think there's a lot of, I would be hard just trying to take, we like treating problems as a healthcare, our healthcare system tends to treat problems after they've already emerged, and at that point in time it's easier to throw medications and other things that people rather than make those lifestyle changes, which I think is quite a shame. And there's a comment which I'll read from Judy Betel again, see the Canadian Center on Activity and Aging, we need to train all CSWs and all staff being trained in supporting this. So another question was code morbidity status taken into account. We did adjust for the number of chronic conditions that people had. And we did the- What chronic conditions, systems? It was the systems, so we were less, if people had multiple very similar conditions, those were clustered, or they're grouped together, so we had musculoskeletal conditions, respiratory conditions, cardiovascular related ones, metabolic, so I think we had 10 categories overall that we looked at for the chronic conditions. Okay. Any final questions to put in? One final comment from Judy. Yes, we give people walkers and wheelchairs for fall prevention rather than exercise prescription. So the need for this research and the communication on the discussion around this research, I think we all agree is very important. Well, thank you again, Alex. Mayhew, we greatly appreciate your participation in the CLSA webinar series and thought it was extremely interesting and good presentation. Okay, well, thank you very much for having me. I'd like to remind everyone that CLSA Data Access request applications are ongoing. Please visit our CLSA website under Data Access to review available data for their information on the CLSA platform and details about the application process. I'd also like to say that our next webinar is scheduled for March. We'll be welcoming Dr. Gilco, Ishe Gama Doyle to talk about assisted devices used among community-dwelling older adults. So please join us for March's CLSA webinar series. So please register soon and join us for that webinar. And thank you for everybody for attending today's presentation.