 longitudinal study on aging to understand the role of mobility testing and fall risk assessment for community dwelling adults. I'd like to introduce our speakers today, Dr. Marla Bochamp and Dr. Ayeshae Quispinar, and just quickly a little bit about them. Dr. Bochamp is a physical therapist and assistant professor in the School of Rehab Science at McMaster. Her research focuses on rehabilitation strategies to enhance mobility among older adults, as well as those with chronic disease. She is particularly interested in advancing evidence-based practice and fall risk assessment and prevention in older adults. And Dr. Ayeshae Quispinar is a physiotherapist and assistant professor in the School of Rehab Science at McMaster. Her research focuses on monitoring health outcomes that are important to people with chronic diseases and older adults including symptoms, function and quality of life, the development of new tools using modern measurement methods, as well as reliability and validity testing of existing tools. So if no further ado, I will turn it over to them. So thanks very much, Jennifer, for causing us. We're happy to be here today. So we are talking today about the general mobility test for fall risk assessment and community dwelling older adults. We'll start with some background on falls and the importance of mobility and balance for fall risk prediction. We'll go through quickly some fall risk assessment and prevention guidelines, and then we'll share with you the results from RCLSA analysis, looking at different mobility tests and their predictive accuracy for falls, and then we'll talk about clinical and research implications of our findings. So why are we here today talking about falls? Falls are arguably one of the most costly and important public health concerns facing older adults. They're the leading cause of injury-related hospitalization among older adults. When an older adult is hospitalized because of a fall, it's associated with the longest length of stay compared to all other causes. Just in direct health care costs alone, falls cost the Canadian health care system $3.3 billion every year. And at an individual level, falls lead to this downward spiral of activity restriction, a further increase in risk of falls, long-term care admission, and mortality. And with the aging population expected to increase, these numbers are projected to increase as well. So given the devastating consequences of falls, it's not surprising that a lot of the literature has been devoted to identifying risk factors for falls. In this very well-known paper by Taneti and Kumar and Jama, we can see that after history of the previous fall, balance impairment is the second most commonly identified risk factor for falling. And the good news is that contrary to what this comic suggests, is that we actually know that falls can be prevented. So there are a number of systematic reviews, meta-analyses, and clinical trials that have shown that if we identify people that are at risk and we provide them with a targeted fall prevention intervention, we can reduce both the rate and the risk of falls by up to 40%. And what types of interventions have the most benefit? Well, consistently, it's exercise programs are emerged as the main component of fall prevention interventions. And in particular, it's functional exercises that challenge balance, that has the greatest impact on falls. So again, the importance of balance and mobility is highlighted here. So fall prevention guidelines have been produced by a number of different organizations. Probably the most common is the AGS-BGS. So that's the guideline put forward by the American and British Diaryatrics Society. There's also the NICE guidelines from the UK National Institute for Health and Care Excellence and the Centers for Disease Control Steady Algorithm Stopping Elderly Accidents, Deaths, and Injuries. And in each of those guidelines, tests of balance and mobility are recommended as part of first level fall risk screening. Trying to advance the slide. So here's a simplified algorithm, which instead of going through each of those guidelines separately, we thought we'd show you a synthesis of the guidelines. So essentially, when a healthcare provider encounters an older adult, and it's defined typically as someone over the age of 65, the healthcare provider should be asking that person, have you had a fall in the last year, or do you feel unsteady with standing or walking? If you answer yes to either of those key questions, if the patient answers yes to either of those key questions, then the health provider is recommended to conduct a balance and mobility screening test to determine whether or not that participant has a high risk of falls. If performance on that screening test is above the cutoff or deemed to not be impaired, the person is considered to either be at lower risk or medium risk, and they receive education and referral to community exercise. If performance on that balance screening test is below the cutoff or deemed to be impaired, then a more detailed balance assessment and assessment of other fall risk factors should follow, followed by targeted fall prevention interventions. So looking at this guideline, one of the questions that is raised is really, well, which test do we use to do the balance and mobility screening test and at what cutoff? And one consideration is certainly that we need to have short, easy to administer tests that make sense in terms to be used for screening. And when we look at the different clinical practice guidelines, a number of different tests are suggested by each, but only one of them actually includes cutoff values to identify people that are impaired. So in the CDC study algorithm, a 12-second cutoff on the time death and go is recommended to identify people at risk of falling. In that same guideline, cutoffs are also suggested for the optional standing balance test and tear stand test that are also recommended in this guideline. However, a problem with these endorsed cutoffs is that there's really limited research to support using them. We often, the study sample sizes are small or based on a convenient sample. And in some cases, they're based on studies where falls are not the primary outcome. So the aim of our study, our analysis today was to determine accuracy and cutoff values of commonly used screening tests of balance and mobility for predicting falls and community dwelling older adults who are enrolled in the Canadian longitudinal study on aging. To answer a research question, we used the CLSA comprehensive cohort baseline data. And the CLSA is really the largest research platform of its kind in Canada, and which follows 50,000 people between the ages of 45 to 85 at baseline over a 20-year period. So for our study for this project, we performed a secondary analysis of the baseline and the 18-month follow-up data from the CLSA comprehensive cohort. And to be in line with clinical practice guidelines, our inclusion criteria were that participants had to be older than 65 years old. They had to report an injury due to a fall in the past 12 months at baseline or report difficulty with mobility during activities of daily living, such as walking, transferring, mobility around the community, shopping, or housework. The CLSA falls at baseline is captured by first asking participants the following question. So in the past 12 months, have you had any injuries that were serious enough to limit some of your normal activities, for example, a broken bone or a bad cut or burn? Participants who answer yes to this question are then asked, was this injury caused by a fall? So those who answer yes, this injury was caused by a fall, and whether the fall was from the same level or from a height were recorded as having a fall at baseline in our analysis. It was falls at approximately 18 months after baseline assessment, which was collected through the maintaining contact questionnaire that's administered by phone. So for the follow up question, participants were asked, we are interested in falls where you hurt yourself enough to limit some of your normal activity. In the past 12 months, did you have any falls? If they answered yes to this question, they were recorded as having any fall in our analysis, so one or more falls. And a subsequent question was, how many times have you fallen in the past 12 months? And we recorded the number of times the fall occurred was recorded, and if it was two or more times, this was recorded as multiple falls in our analysis. Closure variables were poor mobility and balance tests found in the comprehensive cohort at baseline. So we had the timed up and go or tug, standing balance test, which is also known as the single leg stance test, chair rise test, and skate speed. I'm going to go through each one in a bit more detail in the next slide. Timed up and go test or the tug involves participants getting up from a chair, walking three meters, turning around, walking back, and sitting down in the chair at their usual pace. It is really the most widely suggested test by clinical practice guidelines, but the problem is that the data to support the recommended cutoff, as Marla mentioned earlier, of 12 seconds is weak. The 12 second cutoff was actually based on a study that looked at the ability of the tug to discriminate between community dwelling and institutionalized older adults rather than falls as an outcome. Although other studies have validated this cutoff for falls with varying degrees of accuracy, a recent systematic review that specifically wanted to look at the predictive validity of the tug for predicting falls in community dwelling older adults showed inconsistent results and where accuracy or area under the curve values ranged from 0.6 to 0.7. Many of the studies were limited by their small sample sizes. The next test that we used in our analysis was standing balance test. This asks participants are asked to stand on one leg for up to 60 seconds, and the maximum amount of time that they're able to hold that position is recorded. Being able to hold that tandem or single leg standing position for less than 10 seconds is suggested as the cutoff value for fall risk. Again, there are actually no prospective cohort studies that have really looked at this threshold value, and the studies that have assessed it have reported low accuracy with area under the curve values of less than 0.6 test is the chair rise test. This involves participants standing up from a chair then sitting back down five times, and the time it takes to complete this task is recorded. There's some evidence for the predictive validity of this test for falls, but its accuracy has not yet been reported, and as a result of this, there are no established cutoff values or AUC values reported to date on this test. Last but not least definitely is the gate speed test in the CLSA, which is actually referred to as the timed four meter walk test in the CLSA. For this, participants are instructed to walk at their usual pace until they pass the four meter finish line located just four meters away, and that in fact studies to date have shown that their inconsistent results for predicting falls, and one study interestingly showed that the relationship between gate speed and falls may actually be non-linear. Analysis, we looked at the area under the curve of the receiver operating characteristic curve to analyze their data, and the AUC allowed us to really determine the accuracy of each of these screening tests in terms of how well they could distinguish between fallers and non-fallers. The optimal cutoff value was selected based on maximizing sensitivity and specificity at 18 months, and our outcomes were any fall, so one or more fall, or multiple falls, so two or more falls, and we considered an AUC value of .7 to be acceptable. Okay, so this flow chart here shows how our sample was selected. So the comprehensive cohort includes 12,646 people who are over the age of 65 years old at baseline, and among this cohort we almost 11,000 older adults had no falls or mobility limitations at baseline based on our criteria and the questions we looked at, which left 1,719 older adults for inclusion, but we wanted to ensure that there was at least 12 months of time interval between the baseline assessment and the follow-up question or the maintaining contact question at approximately which should be around 18 months. So if there were individuals who were assessed for the maintaining contact questionnaire before this 12-month time period, we excluded those individuals. This left us about 1,121 participants for inclusion in our analysis, and among these individuals, 419 reported having a baseline fall, 646 reported having some form of mobility limitation based on walking and the questions from the activities of daily living questionnaire, and 56 individuals reported having both a fall and a mobility limitation. And if we look at and hear interestingly for each of these groups, there were more women ranging from 58 to 72 percent than men, which was about from 30 to 42 percent. Okay, so shown in this table, so our total sample was 1,121 individuals and of which 67 percent, so 747 were women, while we had only 374 men that were included in our study. In general, we see a trend here that more women had four or more chronic conditions than men, and depressive symptoms on average, we could see that more women reported having depressive symptoms than men with lower education levels and lower income compared to men. In addition, we also can see here that more women reported moderate or severe pain and slightly higher med use for depression. For the baseline performance on the test, our results were actually consistent with what we would expect in a community dwelling older adult sample. Our mean was around 11 seconds for the timed up and go test, about 0.86 to 0.88 meters per second for gait speed. Balance was ranged from 24 to 28 seconds, and chair rise was about 15 seconds, and we could see that as expected in those who were 75 years and older, performance on these tests worsened, and in addition, some of these tests showed more of a difference between men and women. So for example, if we look at the single leg dance test in women, we see on average about nine seconds, while in men it was 15 seconds, outcome at 18 months, which one of our outcomes, which was any fall in the past 12 months. If we look at the slide here, we see that the rates in this sample were fairly similar across men and women, about 19 to 20 percent, which makes sense given that these individuals all had either a previous fall or mobility problem. And in this slide here, we're looking at multiple falls. So our other outcome at 18 months, and 58 out of 747 women, so 7.8 percent reported two or more falls, and 26 out of 374 or 7 percent of men reported two or more falls. And because we had less men in our samples in total, you can see that the absolute number of falls for men by age strata was small, with only 12 repeated falls in 65 plus and 14 in the 75 plus group for men. Current guidelines, our first step was to really analyze our data or our findings with men and women combined and all older adults above the 65-year-old together. So we found that, and as we can see on the table here, that none of the tests actually predicted one or more falls with acceptable accuracy. And our AUC values ranged from 0.5, which is no better than chance, to 0.6 for the tug, which that was the highest AUC value or area under the curve value that we were able to see. And in this table, where when we looked at multiple falls or repeat fallers, the accuracy of the test improved slightly, with the tug achieving an area under the curve value of 0.68 and gate speed falling close behind at 0.65. And here are the ROC curves for multiple falls. So the tug had identified multiple fallers with a cutoff of, we have a cutoff or threshold value of 13.7 seconds. And for gate speed, we have the cutoff of 0.73 meters per second. When we look at the entire sample together. So next we conducted age and sex stratified analyses, and we were able to do that because of the large sample size in the CLSA. So if we start by looking at women, and we're looking at, in this table, predictive accuracy for any falls, these are women that reported one or more falls. On the top, we see women age 65 to 74 years old, and on the bottom are for the women 75 plus. And we can see the AUC values again, similar to when we combine the sample together, the AUC values are fairly low, with the highest AUC of 0.6 being achieved by the time death and go. And the same in the 75 plus category. What is interesting to note is when we look at the model that includes other fall risk factors, so the models that include depression, cognition, vision, education, pain, use of psychotropic medications. When we look at that model, the AUC has actually improved to acceptable levels. So this suggests that when thinking about predicting one or more falls in women, we may need to think about other risk factors. Now looking at the predictive accuracy for multiple falls in women, so these are women that reported two or more falls of 18 months follow-up, we see a little bit of a different story. So now in women age 65 to 74 years old, the time death and go now achieves our acceptable accuracy level of 0.70 in terms of the AUC. The sensitivity at the cutoff value that optimized sensitivity and specificity was 52%, so a little bit low, and specificity was 88%. The positive predictive value was 29%, and the negative predictive value was 95%. So the tests are really able, this test seems to be able to rule out people better than rule them in. In terms of looking at the model that includes other risk factors, again the AUCs improve when we include other risk factors in the model. Looking at women over the age of 75, again the time death and go had the best accuracy with an AUC value of 0.7. Here the sensitivity has improved to 70%, at a slight cost to specificity at 64%, the positive predictive value was 14%, and the negative predictive value was 96% here. It's interesting also the gate speed, although it doesn't reach our cutoff of 0.7, it did come pretty close with a AUC of 0.68 in women 75+. So these are the ROC curves for the time death and go for identifying multiple fallers in women. On the left-hand side is the curve for women ages age 65 to 74, and on the right-hand side is for those 75+, and so both had the AUC of 0.7. The cutoff score for the time death and go in women age 55 to 74 was 14.1 seconds, and this was the cutoff that maximized both sensitivity and specificity. And the cutoff score in women aged 75 and over was 12.9 seconds. Now looking at the predictive accuracy for one or more falls in men, we see a similar trend to what we had observed in women. So in men 65 to 74 years old, the AUC values in general were below what we would consider to be acceptable, so as little as chance with 0.50 to the highest accuracy in the 65 to 74 category being achieved by actually the single leg balance test in this case. And then in the 75+, we have the time death and go is achieving the best accuracy, but still well below what we would consider acceptable at 0.63. Again here, it's really important to note that when we include other risk factors in the models, our AUC values do improve to what we would consider to be acceptable for screening. Looking at now at the predictive accuracy for multiple falls in men, in this case, we have balance, single leg balance being the test that has a very high AUC with an AUC of 0.85. The cutoff score identified was 3.6 seconds and that had 88% sensitivity and 83% specificity, so pretty good. The positive predictive value was 24% and the negative predictive value was 99%. What's interesting if we keep kind of looking along that row, if we look at the model that has other risk factors included, we are able to get very good if not in some cases near close to perfect discrimination when we add other risk factors. In men 75 and older, none of the tests have high accuracy for identifying fallers or multiple fallers and the confidence intervals are quite wide. What is interesting is again when we look at those models that have the other risk factors, the AUC values in some cases go up to quite high. These are the ROC curves for multiple fallers in men and on the left hand side is the one for single leg stance or the standing balance test. In the middle is the time that can go and on the right is the gate speed and we've included the tug and the gate speed here because they were the other two tests that had the second and third best accuracy for identifying multiple fallers in men. One thing to notice here is that again because we had a smaller number of men that met our inclusion criteria for including inclusion into this sample, the number of fall events was lower and so that's why the curves are not as smooth here as in the other ROC curves that we showed. But the cutoff score that we identified on balance in men was 3.6 seconds. For time that can go it was 11.7 seconds and for gate speed it was 0.85 meters per second. So overall we think that these findings raise a few different, I have some key messages. Firstly, none of the mobility and balance screening tests were able to predict one or more falls of 18 months. This is notable considering that we included within the CLSA we have a lot of the commonly used and commonly recommended tests for doing balance and mobility screening. And this might suggest that we may need to consider different balance tests or different mobility tests that either have a higher difficulty level or include more items that challenge more different aspects of balance and these might have higher accuracy. But of course there is going to be a trade-off between complexity of a test and feasibility for using it and screening. In addition, because our AUC values especially in the Asian sex stratified analyses tended to improve when we considered, when we added other fall risk factors in the model, this suggests we might want to think about considering a fall risk index in future work that takes into account some of these other risk factors. Our results did show that we were able to identify those who were at the highest risk for repeat falls. So we were able to identify those high-risk people with the mobility and balance screening tests. And really importantly the optimal cut-off values and the predictive accuracy for the test were different for men versus women and they were also different across age groups. And we believe this has important implications for fall risk assessment and prevention guidelines which to date haven't really considered sex or age in their recommendations. So getting back to the question, which test is best? Well, the answer is not necessarily just one test. In women, the timed up and go has the best accuracy for predicting multiple falls in our study. And again, the cut-off value was depended a little bit on age. The 65 to 74, the timed up and go score was 14.1 seconds that identified people with multiple falls. And in women age 75 plus, the timed up and go score was 12.9 seconds to identify people with two or more falls. And in men, the single leg balance test really had the best accuracy for predicting multiple falls with a cut-off value of less than 3.6 seconds as identifying those with two or more falls. If we look at the timed up and go score cut-off that identified men with multiple falls, it was 11.7 seconds. And I think it's just really interesting to look at how these timed up and go cut-offs change by looking at, versus looking at women, looking at men, and then looking at across the age groups. And you see that they're all, if we were to go with the 12-second cut-off that's currently recommended by the CDC, it doesn't necessarily capture some of this complexity. So some limitations of our study is certainly that the wording of the fall, the questions around falls and mobility were different than in the fall prevention guidelines. So that's important to note. We did have, as we said, a smaller sample size in men, so fewer fall events. And in particular, more, I think we need more research in men 75 years and older. Because in that group, it appeared that falls could be predicted by factors other than balance and mobility. In addition, it's false individuals who were at the highest risk of falls may have been less likely to attend an in-person assessment at baseline in the CLSA. And participants with worse balance or mobility may have been contraindicated to perform the test or may not have performed the test as part of the study. So we may have some missing data related to that. Conclusions based on our findings are that our results are consistent in showing differences between age, differences by age and by sex in terms of the predictive value of the mobility screening test. And I think this is an important consideration for clinical practice guidelines. Future work should also evaluate other screening tests or a fall risk index that would incorporate other fall risk factors. And prospective studies that have been designed to explicitly answer questions about fall risk screening would be ideal to be able to draw definitive conclusions. And we're excited to be starting one of these studies very shortly. So we just want to acknowledge the co-authors on this project, Dr. Parmentarina, who is the lead PI of the CLSA. So Helnaz Moul, who's our statistician, Lauren Griffiths and Alexandra Mayhew for their support and then the funders for the CLSA. And actually both so Hel and Parmentar are in the room with us today so they may type in to help us answer some questions. So thanks very much. Great. Sort of take over from here and help moderate the question and answer period. Thank you for your excellent presentation. Again, I'd like to open it up to questions. I know a few of you have already posted a few in the chat box. So thank you very much. We'll start with those. Just a reminder that the muting will remain on, but you can enter your questions straight into the chat box that's in the bottom right-hand corner of the WebEx window. So we'll start with the first question. I'll just start from the top and if either Dr. Kosminar or Dr. Poshamp want to just sort of filter the questions on their own, that's fine or else I'll just make sure we address them all. So the first question is related to the CLSA as a whole and that is what is the proportion of women in the CLSA samples as a whole? I don't know. Do you both know that offhand or I could probably speak to it or 51 percent? That's what I thought it was. So thank you for your vendor. So it's just over I think it's exactly 51 or just slightly over that. So it's comparable. The next question is why were the MCQ less than six months and six to 11 months excluded? I think you might have touched on this, but maybe if you can explain that again. Well our outcome was falls at 18 months using the maintaining contact questionnaire and that the question the maintaining contact question asked participants to recall falls that they have had in the last 12 months that were serious enough to limit some of their normal activities. And because we're asking people to think back to a 12-month recall period we wanted to make sure that that question was administered at least 12 months after their baseline assessment. Great. Thanks for clarifying. The next question is one of a few from Ashok. Am I correct in inferring that all your models were stratified by gender and age groups while controlling for demographic variables like education, depression, scores, income, etc? So yeah we looked at the data by just looking at the test on their own. So not controlling or including any other variables in the model. So in our tables on one side you can see that it was just based on what are the cutoff values for these tests on their own for different gender and age groups. And then on the other side of the table we had models that included other demographic variables and we did and those were ones like education, depression, income was not included, pain, and medication used for depression. So we had both. We looked at the data both ways. So the first data that we went through was all the data combined. So looking at people over the age of 65 and men and women combined and then the other analyses we presented were separately for women and men and also stratified by age. And the next question is also about models. Did your models include the sampling weights of the CLSA and if so can you mention which sampling weights were used since I know that the CLSA has various sample weight variables? So fortunately I think we might so Hel might know that answer. So we did not use sampling weights for this analysis but I'm not sure if anyone wants to comment anymore on that. Initially we decided to not to use weight and because it's a sub-sample for this specifically for the baseline polar only that's why we did not test it whether the sample weight will be valued or not for this case. So we definitely we can test it later on to see whether we can use the weight or not. Basically if we would have used this analytical weight specifically for this analysis the way we have sub-selected the sample those weights wouldn't have been valid in order to use the weights we would have to recalculate those weights to come up with a valid analytical weight. And that was the rather than doing that we went with unrelated. Thank you. Okay the next question is was there a test for men aged 75 to 85? So in that age group none of the none of the test had acceptable what we would consider to be acceptable accuracy for identifying either one or more falls or multiple fallers but we did what we I think I think the time that didn't go if I'm remembering stuff. The time that didn't go probably had the best AUC in that group it was statistically significant AUC was 0.63 but when we when we looked at the models adjusted or not adjusted the models that included other fall risk factors those we were able to predict both one or more falls and multiple falls with much better accuracy. So I think in terms of this sample of men over the age of over the age of 75 we probably need to look at tests beyond balance and mobility or and really think about other risk factors and that sort of what had us thinking about you know potentially doing a fall risk index in future work that would incorporate some other risk factors. So the next question is am I correct in inferring that the cutoffs were also a parameter of interest that was estimated in your models? Lots of good questions about the models. Not sure I fully understand the question but the in terms of the models that included the other risk factors we would have included the cutoff value that was identified with the on its own that maximized sensitivity and specificity for the outcome and that cutoff value would have been included in that in the model that has the other risk factors. Is that sorry? Does anyone else have anything to add? I don't think so. Let's see if we can okay next question is pulling in the idea of cognition. So how is cognition accounted for and how do we know that participants were actually able to recall the falls that they had? We don't we don't know we know there is we know that we certainly know that recall bias can be a problem when asking participants to recall over the over a 12 month period and typically that results in an underestimation of the number of falls that people report. In terms of cognition I mean we have adjusted for it using the mental alteration test in our in the not adjusted we have included people that were we have included cognition sorry as a factor in our models that include other fall risk factors but we didn't we did not have any exclusion criteria related to cognition at baseline and in in general I so and if we based on our sample our characteristics that at baseline just looking at the math or mental alteration test and those who were below less than 35 which is the cutoff for some cognitive deficits there were it was about eight to nine percent of our sample. So majority had higher scores on this test but definitely do bring up a good point. We still have another at least another 10 minutes for questions but I just want to remind anyone who does need to leave early to complete the survey that was just should have just popped up on your screens before you go. So thanks thanks for doing that for us. Now we'll just go to the next question which is can you comment on whether underreporting of falls could be occurring and whether it might differ by gender age or other characteristics and also how might this influence the findings? So definitely underreporting of falls was probably likely to have have occurred just because we are asking people to recall back 12 months and studies that have examined recalls recall bias have have shown that it can result in an underestimation of the number of falls and the other thing is is that we the wording of the falls question at the maintaining contact questionnaire is really talking about falls where you hurt yourself enough to limit some of your normal activities. So again here we're asking people to recall serious falls or more major falls so it means that some falls that were less memorable may not have been reported so in general we think that that it would be an underestimation and so obviously if we had had more fall events it's possible that it's possible that our accuracy it's possible that our AUC values were underestimated but it's hard to say for sure. We also just have a kind note to say thank you very much for answering all my questions great presentation and research work thanks to all the authors for this work. I might just open it up to anyone who's here in the room right now just to ask if there's anything else you wanted to add to the presentation or respond to in terms of any of the questions. I think just I think the take home from our results is not that we're saying necessarily that these are the the cutoff values that that we need to be you know that we need to be using now it's the point that we probably need to start thinking about age and sex specific cutoff fall risk assessment criteria because our results do consistently no matter which way and no matter which analysis we looked at do consistently show differences between men and women and also differences by age and I think that's important to consider. So your presentation indicated that functional exercises that challenge balance have the greatest impact on falls there's a question could you give some examples of this? So what we mean by this are oh so basically the exercises need to be as opposed to sitting down in a chair and doing strengthening you have to be up you have to be doing activities that you would do in your daily life that's what we mean by functional so so that means you know walking that means sitting sitting down in a chair and getting back up it means doing tasks that you do in your daily life and using that to train in terms of challenging balance during functional training that means there's a number of different ways we can challenge balance but by narrowing the base of support is one way by adding a dual task is another by making by altering speed by removing visual input those are all ways that we can challenge balance while doing functional exercises. And another question that I had is that obviously there's differences in sex among the results and you also mentioned implications in terms of this in terms of fall risk assessment guidelines I'm wondering if you might be able to give some examples of what the implications of these differences might actually look like in clinical practice in these guidelines. So a bit of a knowledge translation question. I think so in terms of clinical practice I mean what we as a clinician you cut off values and that's why we it was one of our overall findings isn't in terms of looking at as the clinician you would want to know does my patient is my patient over or above or under this cut off value so it can tell you a lot in terms of okay who's coming in to see me what is their current status and then are they above or below this threshold and are they at risk for fall right so and I think our results show that there is a trend that these values are different for men and women and then that we need to assess accordingly and look at cut off values separately for each one. So what it might ultimately look like is just instead of just one cut off value for everybody it might be it might ultimately be you know for men of this age you know we should use this test and at this cut off value and for women of this age we should use this test at this at this cut off value and obviously I'm trying to keep the messaging as simple as possible but making sure that we're you know providing as precise estimates of the of the cut off values as we can. Thank you. Well it sounds like there's lots of opportunities to take this and translate it into useful guidelines for clinicians like yourself whether it be in primary care or wherever the our patients in the community are being seen. So I don't see any other questions so maybe I'll wrap it up at this point. Feel free to post any last questions while I go through the closing of the webinar. So again thank you again for such a great presentation. We appreciate your presentation in the webinar series. I know that you've presented for us before and it's again greatly valued and I know very people are very receptive to this topic. I'd like to remind everyone that CLSA data access request applications are ongoing. The next deadline for applications is February 12th of 2020 which is only a few months away. Please visit the CLSA website under data access to review available data and change further information or details about the application process. I'd also like to remind everyone to complete the survey that's located under the polling option. If you have any questions or concerns that any of the team here at the CLSA can help with please write to us in the chat box and we'd be happy to help. Our final webinar of 2019 will take place on December 16th with Dr. Ruth Barclay. Yeah I wanted to make sure I said that right Barclay. She's an associate professor in the Department of Physical Therapy at the University of Manitoba. Her presentation will be factors associated with community ambulation in older adults and those with stroke and osteoarthritis. So if you're interested in that webinar we encourage you to register as soon as possible. Also graduate students and postdoctoral fellows for any of of you that either have students or postdocs or if you are one and you have an interest in longitudinal studies on aging. We encourage you to save the date for what's called SPA 2020. This is an innovative five-day training program that will take place next June at Hawkely Valley Resort in the southwestern Ontario. More details can be found. Will be available in January 2020 when the program launches on CIHR's research net. And remember the CLSA promotes this webinar series using the hashtag CLSA webinar and we invite you to follow us on Twitter at clsa underscore ELCV. And please go to our CLSA website to register for our webinar series again and join us for upcoming webinars. And finally thank you again for all of you for attending and for our presenters who also presented. Thank you.