 Okay, good afternoon everybody to the CLSA webinar series. My name is Ina Rubin, I'm the Managing Director of the Canadian Longitudinal Study on Aging and you're all welcome at today's webinar. And with that I'm going to start with an introduction of Gillian Mulvail. Gillian is an Assistant Professor of Health Policy and Analysis in the DeGroote School of Business at McMass University. Her research focuses on improving care coordination across health professions, sectors and stages of the lifespan through the development of health policy and management frameworks that promote interprofessional person and family-centered care with applications in mental health and primary healthcare. So Gillian derives theoretical approaches from interdisciplinary training in health policy analysis, health economics and health research methods. So and Gillian is the co-author of the recent report on the prevalence and cost of dementia in Canada, a review of the evidence commissioned by the Alzheimer's Society of Canada and she is going to tell us all about that today. So over to you Gillian. Thanks so much Ina, hopefully everyone can hear me okay. It's a pleasure to be here and I really welcome the opportunity to speak with everyone about the recent report as Ina mentioned. Looking at this important question of how do you go about estimating dementia problems and are there opportunities from the CLSA? And we were really tasked with the challenge of how do we understand the various estimates that are out there. Not so much to come up with the one single estimate, but to really shed light on how those estimates are developed and how the various assumptions being made influence those. So I'm sure there's a background as I'm sure everyone on the line is aware the problems of dementia increases sharply with age and some estimates suggest that 15% of the population seven weeks to 74 years of age would have dementia and that when you get into the age category of 90 plus some estimates suggest it could be as high as 40% of the population. And since we know we have an aging population and demographic projections suggest the population age 70 plus is going to double over the next 20 years. Obviously from policy and health care management perspective this is a really important question. And so there's important implications for the future need and cost for health and long-term care services. And so in light of this situation the Alzheimer's Society of Canada and the Public Health Agency of Canada commissioned a report entitled on the problems and costs of dementia in Canada, a review of the evidence. The lead authors were Michelle Grigno and Byron Spencer from McMaster University and Susan Braunskill from Sunnybrook Research Center and Institute for Clinical Evaluated Sciences and myself are also co-authors on that report. And the goal here was really to provide a critical review of the evidence as it relates to current problems of dementia in Canada and costs associated with dementia. Secondly, to assess the methods used to reject these estimates of problems and costs. And this was really the critical piece to really try to understand these estimates. The focus of the overall report was on Canada and selected comparator countries. Here we're going to focus on the estimates for Canada. I'll touch upon the comparator country estimates, but that won't be the focus of today's webinar. And also evidence based on survey and administrative records. Again, we're going to focus on the survey as the source of different estimates. Also note that there was a whole piece of the report around costs. We won't be touching that today, but it's something if you're interested that you can pursue in the report. So measuring dementia. Of course, it's important that we get a definition that's clear. We drew upon Sheehan's definition that dementia is a clinical syndrome characterized by progressive acquired global impairments of cognitive skills and ability to function independently. And I think there's a piece here, and that's the progressive nature of the illness, which is a real challenge when it comes to measurement because it's important to decide at what point does cognitive impairment meet the criteria for dementia. And then within that broad definition of dementia, there can be a variety of levels of severity. And that's important when you try to measure estimates of prevalence. Are you comparing apples and oranges in terms of the severity of the illness when you're looking across jurisdictions, points of time, different estimates? And so Sheehan also did a review of the various methods we could use to assess and measure test for dementia and identified 10 different methods that include a cognitive test for cognitive impairment. Overall, comments were that most of these don't take very long to do and yet perform well, both in terms of sensitivity and specificity to its own gold standard. Within each study, there would be a gold standard they compare again. However, overall, there is not one clear gold standard measurement for dementia. So measuring prevalence and incidence. The other important thing to consider, and I'm sure many of you are aware of this, but just so that we're all on the same page, when we talk about prevalence and incidence, there's an important distinction. Prevalence is a stock measure. In other words, it estimates either the number of cases at a point in time or the rate in a given population group at a point in time. Instead, or in comparison, incidence is a flow measure. So it's the onset of cases in a given time period or the change in rate of prevalence in a time period. So these measures are not always a one-to-one. If we know prevalence, we know incidence, we can just automatically adjust prevalence because we need to understand the differences in survival for different people with various cases of dementia. It's not uniform. And so rates can be measured for the population as a whole with or without age standardization or they can be age standardized rates. And these are very helpful. What they do is they measure prevalence relative to a specific population age distribution. And so it makes it easier to compare rates across time, changes in rates, or across jurisdictions. Also note that when you're being presented with a prevalence rate, it's really important to understand what the denominator is. What is the age group or sex group that it's referring to? So rates can be measured for the population as a whole. This will be much smaller than when they're estimated for, say, the population 65-plus or the population eight years or older. So there's other challenges and measurements that we also need to consider. First, how do we get reliable measures of new cases of incidence? We know that diagnosis can occur at different stages of severity. So someone may meet the criteria for dementia and have a mild case. Well, another person will have a severe case, and yet they would still both be counted as an additional case of dementia. But we may be comparing very different disease states which have important implications for policy and healthcare delivery. Second, severity itself is difficult to measure. So it's easier to identify a case than to identify the severity of that case. And registration of new cases is not something that's mandated. So it's not like we have a directory or a registration that we can just come up with the numbers of all cases that are out there. So in terms of how do we come up with estimates then? Well, there's two main data sources that we look at. Survey estimates and administrative records. And so the survey estimates can be either self-reported, so a survey comes to a person and that person indicates that, yes, they or a family member has received a diagnosis of dementia from a healthcare professional, or the assessment can be done directly by a health professional. And we have examples of surveys that use each of these approaches, the Canadian Community Health Survey or CCHS, and the National Population Health Survey, both use self-reported data in their basic form. And the Canadian Study on Health and Aging used health assessment by a health professional. Now I'll get into more details about each of those surveys a little bit further into the webinar. Administrative records is another approach. And in brief, what this approach does is it analyzes records in administrative databases. So that could be position billing data, for example. And it uses algorithms that examine those records to say, given certain classes of medication that are largely prescribed or targeted to dementia, how many individuals would meet those criteria, also other risk factors are taken into account in these algorithms, persons age, sex, and so on. And so that's another approach to trying to come up with estimates. So if we think about surveys of data sources versus the administrative data, there's advantages and disadvantages. So of course, advantages of surveys is that they can really be targeted to understanding particularly disease categories. Since the survey is being designed, you can ask the kinds of questions you want. You can focus particularly on different disease categories and go into a lot of depth there. However, it's an expensive thing to do, to do a population level survey. It requires a large sample size. And also, it's difficult often to come up with a survey that will capture prevalence rates both in institutionalized and community populations. Often you need a different approach in either who's answering the questionnaire in an institutional setting, it may be a care provider or a staff person. In the community, it's likely to be the person that solves our family member. And then there's issues around the accuracy of the survey. And that depends very much on how the sample is structured, response rates among participants. And then so accurately, they're able to respond to the survey questions. So there may be some discomfort in acknowledging a diagnosis by a physician of dementia for a person or family member or other biases that can result or perhaps misunderstanding by participants in responding to a survey. And then issues around how large is the sample size, the higher the sample size, the more precise the estimates are going to be. And finally, most surveys are cross-sectional. You can have longitudinal surveys, which are year after year surveys of the same participant. Of course, those would be more expensive. So most of them are cross-sectional. And when you have a cross-sectional survey, what that means is you get a snapshot of problems at a point in time, but you don't get information on when did that person first experience the disease or their incidence, I should say, or leave the disease, in this case, at death. And also it can challenge our ability to do comparisons over time and across countries. If we can think about administrative data sources, they too have their advantages and disadvantages. Some of the advantages of administrative data, typically there's information on hospital stays, doctors visits and prescriptions that may be fairly accurate relative to self-recording in a survey. It can also be used to infer the presence of a particular disease. It's a cheap data source because it's usually being already gathered for other purposes, so there's no incremental cost associated with using this administrative data, or I should say relatively low, there's always some cost. They're usually representative because there's often very, very large samples of sample, most people go to see a family physician every year, and so you'll get a really good sample size in that sense, and there's also the opportunity to link to other data, so while there may not be an awful lot of other data perhaps around lifestyle, sometimes we can link administrative data to other survey sources to gather information. However, there's disadvantages here too. They're not really, these databases aren't designed for this purpose, and so how well they capture and estimate presence of dementia is a question. How well do the algorithms perform? There can also be variation, just as there's issues with self-report in a survey, there can be variation in how healthcare providers provide data to these administrative data sources or billing records, for example. Are they consistent in how they classify information that they pass along? And in addition, there's limited additional information to determine the determinants and precursors of the disease, which you may be able to more fully explore in a survey. As I say, this webinar is gonna focus on our findings with regard to survey estimates, and in order to come up with that information, we conducted a targeted literature review, not a systematic review, where we tried to identify the key surveys, published prevalence rates and projections of dementia prevalence in Canada, and a selected group of comparator jurisdictions. So we focused for comparators on the UK, Europe as a whole, and the United States. We conducted a Medline electronic database search over the period January, 1990 to October, 2014, and we used key words related to dementia, prevalence, surveys, and methods, and variations on those. We also carried out Google Scholar and Google searches to capture gray literature. We also explored government and survey websites to look for other reports that may be relevant, and we did reference checking of all of the various literature that we've websites that we examined. And then it was an iterative search. We went back to the literature for published academic publications on the methods used in producing these various estimates and surveys. All together we identified more than 40 publications of these 26 contained prevalence estimates and or projections. We excluded any articles at this round that were focused exclusively on methodological issues. So they came in in the second iteration of the search, or though that reported prevalence rates for countries that weren't on our list of comparators. So clearly this is not a systematic review of every estimate out there, rather a targeted review that reports on what were felt to be the most relevant results for our purposes. So just to give you a bit of an overview of the various national surveys that we found that have a dimension of focus. First, the Canadian Longitudinal Study on Aging Neurological Conditions Initiative. As many of you are aware, this is the developing research program, and so the results in terms of estimates are not yet available. The Canadian Community Health Survey for the cycles in 2010-11, there was a Neurological Conditions Module that was added to the traditional CCHS survey. With a follow-up survey, if the respondents or household members showed or indicated a presence of Alzheimer's or other selective neurological disorders. Third survey was done within institutions, institutional settings, so CCHS is a community setting. Here are the survey of neurological conditions in Canada for 2011 and 2012. And this was a survey done by mail, from mailed out to participants who were long-term care staff, and they were asked about the prevalence of 50 neurological conditions among their residents. There were also some other surveys that we identified, national surveys that were not targeted for producing estimates of prevalence of dementia, but nonetheless have been used for that. And those were the Canadian study on health and aging, the CSHA, and this was a broader study that looked at the health and aging population in general. Within that, there was a significant piece around estimating dementia. And so 1991, 1996, and 2001 were the years of data collection. Also, the National Population Health Survey, which is an ongoing longitudinal survey, whose primary focus is not dementia, but a broad range of health conditions, and the Canadian Community Health Survey, which is a cross-sectional survey, and longitudinal components has been added every other year from 2001 to 2007, and annually since 2008. Again, looking at the broader health of the population and health services utilization, other measures not uniquely focused on dementia. The Canadian study on health and aging, this study in particular has a multi-step neuro-psychological and clinical assessment to really try to accurately capture dementia among participants. The prevalence estimates based on this survey align with the DSM-3, and the core objectives of this survey were to measure incidence and progression of severity of the disease, which requires a longitudinal approach to see how participants are moving over time, so we need these repeated observations on the same sample. And this provides data on the progress of dementia across various stages of severity, starting with dementia at baseline, and for those who don't have dementia at baseline, how do we see over time then progress towards dementia? Comparing this study with the NPHS of the National Population Health Survey and the Canadian Community Health Survey, they have collected dementia data by self-report. So in this case, the person being surveyed or answering the questionnaire would be asked, has a physician ever diagnosed Alzheimer's disease or any other dementia for you or a household member? And they're responding to that question. And the overarching objectives of these surveys are really not, as I mentioned earlier, to understand dementia in particular, but the broader health of the population and its determinants to come up with regional and national estimates. And so it can be useful, but it was not intended for coming up with estimates of dementia problems. And finally, a little bit more detail on the survey of neurological conditions in Canada. As I mentioned, it was a one-time national level survey of long-term care facilities. Questionnaires completed by staff of those facilities. And they indicated the number of residents in total and by sex who have been diagnosed with either Alzheimer's disease, any other dementia and 14 other neurological conditions. Again, by mail survey. And one of the limitations here is that it was very targeted and gave very little associated data to understand covariates and other determinants in circumstances. So this slide, I test to give you a bit of an overview of the various surveys in Canada. So the Canadian Study on Health and Aging, National Population Health Survey has three components. A household, that's the HH component, an institutional component, and one for the North. The CCHS and the institutional survey we just discussed as well, the SNCIC. So you can see when you start to compare surveys, there's a lot of dimensions you need to think about. The year in which it was conducted, the design, was it a longitudinal list, the L or a cross-sectional, or both CS, the setting in which it's carried out. Was it carried out in the community setting or an institutional setting or both? You can see here that only the CSHA was in both. The population that's being surveyed. So CSHA focused on the population 65 and older, whereas the community surveys are focusing on 12 years old and older. Sample size, and not surprisingly, the institutional component has a much smaller sample size. If you can see the CCHS, very large sample size, very broad across the country. The method of the interview, or method of data collection, is it an interview with an examination by a physician? Is it by phone? Is it by in-person or is it by mail survey? And then considerable differences in response rates, although they're all fairly good, but something to take in mind when you're considering the accuracy of the survey, as well as how a case is ascertained or determined cases of dementia. And so CSHA is both a self-assessment or a self-screen and then an assessment by a physician, all of the others were self-reports based on an earlier physician diagnosis. So this slide is looking to provide you with an overview of what we came up with in terms of the prevalence estimates for Canada based on these various surveys. And again, you can see that it's difficult to compare and there's lots of spotty places in terms of being able to compare all of these across the board. So depending on the author, we can have different estimates presented. So for the Canadian Study on Health and Aging 1991 data, if we look at the institutional setting, we have differences from 56.9% from the working group, that's the WGE, the CSHA working group, compared with Graham and it was very difficult. We were not able to really ascertain what the differences that caused those different estimates. But there's some general things we can see here also going on. We can see differences in the population over which these estimates are being generated, differences in the year, and then the setting. Is this estimate overall for the whole population within a particular age group at a point in time? Or is it just for community or just for institution? And we see that only the Canadian Study on Health and Aging gave us an overall for both community and institutional. We also see, and that tends to be in the seven to 8% range, we also see that in the community setting, much lower estimates, so 4.2% from the Canadian Study on Health and Aging, 2% CCHS 2003 for the population 65 plus. We see that go down to 1% for the population 55 plus and so on. So depending, it's up to 4.3 when it's 80, age 80 plus and that makes sense. We expect a higher rate in that different population age group and older population age group. We also see much higher rates in the institutional setting, which is not surprising when we think that about severity of illness. And so we see that any number that's out there that you look at, you're gonna have to think carefully about some of these various determinants of that estimate. On this slide, we look at the data from the Canadian Study on Health and Aging in terms of prevalence by setting and severity of illness among those 65 and older. So within the community, institutional setting and total estimates were provided, here we see that in the community setting, the rates are higher for people with mild dementia and lower for severe. Not surprising, see the opposite, where in the institutional setting, we see 31% have severe dementia, 6.6% of those with mild dementia are in the institutional setting. And so overall, approximately half of those with dementia were living in an institutional setting and this varies by level of severity, 85.4% of those diagnosed with severe dementia were in the institutional setting, compared with 20% of those with mild dementia. So turning now to prevalence projections for Canada, what we found in our review was that the projections often can be based on the same survey estimates, can really differ quite significantly and this largely reflects approach and underlying assumptions. And so there were two main approaches that we identified. The first is where age specific prevalence estimates are assumed to remain constant over time, over the projection period. And so we keep the prevalence rates the same in each age category and forecast out or project out what the population size is going to be in each of those groupings. And so when we do that, we're either assuming that there's going to be no change in prevalence because the incidence rate and the mean length of survival are assumed to be unchanged or any changes in those two variables are offsetting each other and balancing each other out. So the second approach starts with base period prevalence rates and each year adjusts that for predicted changes in number of new cases, so the incidence and the number of deaths or mortality. So two very different approaches. And this graph shows us the prevalence of dementia projections for Canada based on three estimates that we were able to come up with in the literature review. So when we look at these, the blue line and the orange line are both being driven by base prevalence rates from the Canadian study on health and aging. And so one set of estimates is from the CSHA working group and the other was produced by Alzheimer's Society in the Rising Tide report. And so they both began with the CSHA data for 1991 but they take different approaches. The working group assumes that constant age specific prevalence rates and so that's the blue line that's not rising as quickly. And then the Rising Tide uses this forward projection method where it takes estimates of new cases and subtracts out mortality and projects a much higher rate of increase in the prevalence rate. Denton and Spencer, which is the gray line at the bottom much lower, it has a different starting point because it's based on the CCHS, the Canadian Community Health Survey prevalence rates. So it's growing off that lower community based prevalence rate in the first place. And then a few of those prevalence rates by age category stay the same and projects out the population growth for the community based population. And we see much lower increase here. So we aren't going to comment on which of these are right, it's something for you to think about. But in terms of our key findings overall for Canada, the overall prevalence rate for the population aged 65 and older in Canada appears to be in the seven to 8% range based on these findings. The Canadian Study on Health and Aging remains the gold standard for survey based estimates of prevalence and dementia in Canada, although we definitely recognize the information is now quite dated. We say this because of the quality of the algorithm, the large sample size and the fact that it captures both the community and the institutional setting. And so what was also I think somewhat reassuring was that even though the approach seems less rigorous in terms of the case ascertainment, similar prevalence estimates are found using data from the National Population Health Survey and the CCHS, which suggests that if you have a really good baseline prevalence estimate from a survey like the CSHA, these ongoing community based surveys can be helpful to see how does that change over time because there seems to be some overall consistency even though it's a less expensive approach. In terms of what we found from the comparative jurisdictions with regard to prevalence rates, I mentioned I'm not gonna go into a lot of detail here or we'd be on the webinar all day, but just to give you a bit of a flavor, the age standardized rates across Canada and Europe seem to be roughly similar in the range of 6 to 8%. So in general, the European results and the UK results are quite similar to those in Canada. However, the US estimates seem to be much higher than all of those countries. For example, for Alzheimer's disease, the estimates are 11 to 13% prevalence rate. And we were not able to come up with a good understanding of why those US estimates are higher. However, we suspect that it's unlikely, that it's strictly because of differences in risk factors and more likely reflects differences in how dementia is defined and diagnosed. Or it could be that people are going to specialist care more quickly and being captured more quickly and diagnosed more quickly. So I say that's not clear to us, but these are some possible explanations. So looking at the comparative jurisdictions in terms of methods, the Delphi consensus process was another approach that we saw being used in the UK and in several European studies to estimate prevalence. So this was neither strictly a survey nor administrative data. And what this approach involved was looking at published survey-based prevalence estimates, doing systematic reviews of those published estimates and then bringing together expert panels and pooled analysis to try to come up with an overall estimate, particularly in Europe across many countries. The importance of definitions was a clear message that came out of reviewing these studies. So comparative analysis, for example, of UK prevalence estimates over time, was there was a really interesting study by Matthews that suggests that we need to really keep a handle on these overall estimates coming out of the large survey. So it's fine to do these Delphi consensus approaches, but every now and then we do need to gather new data periodically because this study suggested that over time we've seen a significant difference in prevalence estimates and the decline over time in the UK from, let's get the numbers here, from 8.3% in 1991 to 6.5% estimate in 2008. So this may be suggesting that there's a later onset, people are healthier, unclear, but this study really said, you know, it's not enough to just take one estimate and keep it the same over time. We need to take a look at what's going on. It also suggests that in order to understand changes over time, we really need consistent definitions of dementia and the various diagnostic criteria being used. And finally, another message that came true was that we really need to make sure that the exclusiveness of the definition is matched to the purpose of the estimate being produced. So if we're developing programs where we need to really give people assistance with daily living, we aren't so much interested in milder cases of dementia. It's probably the more severe cases. And so we need to make sure that we're matching the estimates to the programming needs. So some of the overall findings that we came across in survey-based estimates. Longitudinal studies emerged as a very important data source in all of the jurisdictions we looked at can either be nationally representative standalone projects like the Canadian Study on Health and Aging or smaller scale studies were used both in Canada and the US which draw upon infrastructure of larger nationally representative studies with broader objectives. Canadian Long Student Study on Aging is looking beyond strictly dementia, but that's an example there or the CCHS or National Population Health Survey. And comparing problems rates based on surveys is challenging due to methodological differences, particularly different age bands and subtypes of dementia. So what about the Canadian Longitudinal Study on Aging and the Neurological Conditions Initiative? Well, the primary objective is to support ongoing population-based research on dementia and selective neurological conditions. One of the challenges we have though when it comes to trying to develop problems estimates is that at the baseline, cognitive impairment is an exclusion criteria. So that means that we're unable to come up with estimates of problems of Alzheimer's disease and dementia because at baseline, we're excluding people with cognitive impairment. However, participants can enter the study and become cognitively impaired during the course of the study and they will remain in the study. This means it can be used for incidents and estimates of incidents. The data, there'll be baseline data from a tracking sample, the first 20,000 participants will be asked about Alzheimer's disease and other dementias as a self-report. So did a physician provide a diagnosis for this for you or a family member? However, we know that's going to be an underestimate because of the exclusion criteria of cognitive impairment at baseline. The in-person survey will come from their more comprehensive sample of 30,000 respondents and here we have a case after team algorithm which I'm just gonna go into on the next slide. So this algorithm was developed by the CLS-18 and incorporates a rigorous neuropsychological battery of tests that will develop composite scores in three domains, memory, executive function and psychomotor speed and that will enable the classification of participants into five different categories, normal, multiple-domain cognitive impairments but not dementia and music, cognitive impairment, single-domain non-memory and dementia. And this becomes important because while the CLSA can't come up with an estimate of prevalence of dementia, it is nonetheless a longitudinal survey and that means the baseline survey will give us some information around vulnerability to dementia and other risk factors and then future ways of the survey will provide estimates of incidents. So new cases and increasing severity is something that it can capture as well. Using this categorized data from several ways may assist us down the road in understanding a progression of symptoms toward dementia over time and enable us to generate transition probabilities across stages of the disorder which will be very helpful for programming and policy objectives. So I'm gonna wrap it up there and say thank you very much to the Alzheimer's Society of Canada and the Public Health Agency of Canada for providing a funding for this study and be happy to take questions and foster discussion among all of you. Okay, so I'll start with the first question to Gillian. And in the meantime, I was quite interested to see why the estimates in the US were so much higher and I know you were speaking about the possible explanations of why that is the case because people get diagnosed earlier but you would think that perhaps then the estimate in Canada would be similar to the estimates in the US. I know that you gave those explanations why you didn't really know but might there also be impact of lifestyle for example and do you know if anybody has ever looked at that and when you think about the US numbers? Well, we've spent some time as a group trying to digest why these would be different. And certainly lifestyle can be part of it. When we look at the estimates from the various studies though it really wasn't clear but I do wonder about specialists in the US being more direct than whether that's an incident to be explored and but honestly it ended up being quite a challenging way for us to try to understand that as a research team. And we've recently presented this information to a number of invested individuals and sort of nobody came up with a better explanation but I'd be certainly interested to hear from other participants that they have other ideas as well. Okay, great, thank you. There's another question for you, Jillian, from Christina. She writes, do you have a concern about to use a self report for estimating the problem? Well, I think in any situation, yes, there's going to be challenges with self-supported data. We know that from survey methodology in general that there's always going to be some questions. Dementia is particularly challenging if we may have difficulties with cognition in the first place, self-report can be challenging. And there's also concerns around social sensitivities with regard to reporting dementia on behalf of a family member and so on. So there's a challenge there and yet we were really struck by the fact when we looked at prevalence estimates from the self-reported data from the surveys and the CCHS and the NPHS that we still were seeing some, at the aggregate level, relatively comparable estimates. And so while there's a concern, we also have concerns with the administrative data. And so we have to work with what we've got. It was somewhat reassuring that self-report from the survey data seemed to perform well given its limitations when we compared that with the Canadian study on health and aging estimates. Okay, great. Another question was answered by Andrew and he asked, what about differences in incidence and prevalence by ethnicity? Is that something that your report has looked at? No, it wasn't part of our mandate to look at that by ethnicity. It's certainly an interesting question and an important one that we need to look at down the road to really get a good picture at this but we were really tasked with kind of more of that overall estimates and comparing across countries. I think part of the motivation here was this dramatic difference as we see in the various forecasts of overall prevalence rates even within Canada, but when you start comparing to other countries and when it comes to health policy and management perspective, these are really, really significant for planning purposes. And so our goal was more to kind of understand the estimates that they said than to come up with one particular estimate or to break that down by ethnicity, but it's certainly something that should be explored down the road. Interesting. You also showed some of the estimates for the European countries. Have you looked at any of the estimates for prevalence of dementia, I'd say, in the Eastern countries like India or Japan that is of interest also when it comes to ethnicity in our own country? Is that something that you looked at as well or are you aware of? I don't know. We didn't explore those, you know, as I say, that was beyond the scope of this particular project. But certainly in future research would be very much of interest as you say. It could help to inform some of the questions around the ethnicity for Canada, but also recognizing some of the differences in how these surveys estimates are being captured across countries as well would also need to be taken into consideration. Okay, question by Christina and it says the CSHA found a high prevalence of cognitive impairment but not dementia and can you Chilean comment on how this could be distinguished from dementia in a self report? Sorry, I don't have an answer, but I think, you know, it could very much depend on the kinds of questions that are being asked to try to distinguish that further in terms of, you know, what is the root cause of the cognitive impairment? I would think that the survey questions could be expanded to include that and to understand that a little bit better. Okay, great. So I don't think I see any more questions and we're getting close to it in time. So unless somebody has another question and you can try really fast and we can answer Chilean, ask Chilean these questions. And then eventually I really want to thank you for this great overview of all the different data that's available to determine dementia prevalence in Canada and beyond and thank you very much for your insights. And I also want to thank our participants for logging on and participating in our webinar. So thank you very much. Thank you.