 So, today's webinar, again, the impact of the COVID-19 pandemic on the mental health of older adults. A longitudinal analysis from the CLSA is being presented by Dr. Perman de Reina. Dr. Reina is the lead principal investigator of the CLSA, the scientific director of the McMaster Institute for Research on Aging. And a professor in the Department of Health Research Methods, Evidence and Impact at McMaster. He holds a tier 1, Canada research chair in gyro science and the Raymond and Margaret Labarge chair in research and knowledge application for optimal aging. He specializes in the epidemiology of aging with emphasis on developing the interdisciplinary field of gyro science to better understand aging from cell to society. So, without further ado, I will pass it on to off to Perman de Reina. Great. Thanks, Jennifer. Charlie, do I have the control now? Yeah, you're good to go. Thank you. Well, thank you everyone for joining us this afternoon to hear what I have to say. And as Jennifer introduced that, I'm going to be talking about some of the data we have been analyzing from the CLSA COVID questionnaire data. And I'm here on behalf of the main team members who actually designed and helped launch the CLSA COVID-19 questionnaire study plus who are also members of the Canadian longitudinal study on aging. And so I'm speaking on behalf of them and I couldn't say more or enough about our participants and we have incredible group of CLSA participants who have contributed to all aspects of the CLSA and especially that the CLSA COVID-19 study. And how quickly we launched this study and what type of responses we got is really something to be admired and we are very thankful to their participation. And none of this would have been done without our staff across the country through these very difficult challenging times working from home, dealing with parents and dealing with their children at the same time. So our full thank you and appreciation goes to all of our staff across the country. So, and I couldn't have done or what I'm presenting here without the right and left shoulders that sort of Susan and Tina sort of work with me and we sort of collectively take care of the CLSA. And they were also the main architects of the design of the CLSA COVID study, along with many of these investigators were part of the CLSA team. And I have specifically highlighted some of the people who spent a lot of time with us in designing, developing and solving problems as they emerged with the CLSA COVID questionnaire. And these people are highlighted in an in bold here on this slide. So, before I talk about the CLSA COVID, I think some of you might not, you might be on the call who don't know the CLSA platform and how this study itself was embedded within the CLSA platform. It's important for you to have a sense of what we are doing with core CLSA. So the core CLSA is developed as a research platform to generate data and evidence that informs not only. Not only at the public health policy, but policies at all level and advances the science of aging in this country and around the globe. Just a quick overview. It is a cohort at the baseline was 50,000 people between the ages of 45 and 85 and 20,000 of these were providing data through questionnaires only and they were a random sample of the Canadian population in 10 provinces. And 30,000 is what we call a comprehensive. This is where we do much more in depth investigations. We go to the homes of the people we collect blood samples do we do physical assessments at data collection. And sites and they were all, they are all established from 25 to 50 kilometer of 11 sites across seven provinces. So I won't spend a whole lot of time on the CLSA CLSA design core, but what I'm going to be talking later on, it will become relevant why I needed to show this slide to you. And, and just to put it in the context, they both are comprehensive and the CLSA that they roughly started in 2011 2012, and we finished the first baseline data collection and recruitment by 2015 from 2015 to 2018 we completed our follow up one. And 2018 we started our follow up to, and when we were in 2020 March, we actually suspended our data collection because of the, because of the pandemic, it was right at the beginning, even the lockdowns in most provinces started at on March 16. We had suspended the face to face data collection of the follow up to CLSA on March 8. We pivoted and rather than just collecting no data, we set up all our staff in their homes to collect as much data as we could through telephone interviews. And this is what actually was happening when we were in the middle of the follow up to this is a, it just gives you a sense of what was going on in Canada in early 2020 in March, February to March to April. And that's the first peak, and that's where many of the public health measures, lockdowns, different strategies were being implemented across the country. And a lot of that that was being done not knowing what was coming in the coming months. And then the summer months looked like they were peaks were going down and some of the areas across the country started to relax some of the lockdowns or other public health measures. And then again, as the fall started to emerge more closer to October, November, we started to see the peak for the pandemic again happen and again. All of the restrictions, all of the public health measures, people were staying indoors in homes and it had disproportionately affected at least for the first and the beginning of the second wave. For the elderly people, especially in the long term care, but also in the community also. So that sort of, as this was happening, and we had suspended the CLSA core face to face data collection. Basically, within a week of suspending the face to face data collection we started to discuss whether we should be launching a CLSA COVID-19 questionnaire study. And we were fortunate to get funding from Public Health Agency of Canada. They were, they are one of the major funders along with Jovinsky Research Institute for McMaster. McMaster Provost and Vice President Research Fund and McMaster Institute for Research and Aging. So they are the main funders of this CLSA COVID-19 questionnaire study. But we also got additional funding for the Nova Scotia component from the province of Nova Scotia, Susan Kirkland had applied for these funds and we were able to receive some funds to do some activities in that province. So we launched this study on April 15th to 2020. We actually, if I recall, first time we talked about doing something like this was, I was looking at my notes, it was actually on March 28th. So within roughly two weeks we had driven our staff crazy to develop software, develop questionnaires, program questionnaires and get it out in the field because it was important to capture people's perception. And originally we started this questionnaire to understand the epidemiology of the infection and the psychosocial and other behavioral aspects that people might experience. And so how did we recruit people? We have the core CLSA sample size is little over 51,000 and over from baseline to follow-up one and in the middle of the follow-up two. We've had some 8,638 people who were excluded because of either they had disease, they require a proxy, either they have withdrawn or for some other administrative reasons we were not able to contact them. To make the long story short, our eligible sample for the COVID-19 questionnaire study was roughly around 42,000 people. And we got a response rate of close to 67%, 28,559 responded to our baseline COVID questionnaire. And this is what the design of the study looked like. So large proportion of people in the CLSA were able to complete the questionnaires on the website that the web based interviews. But there were around 20% or 30% of the participants who didn't have access to internet, we collected the same data using phones. So the both sides had the baseline interviews, for logistic reasons, the phone based interviews were then repeated questionnaires were, data were collected by weekly. And in the baseline originally we started with weekly data collection but quickly we realized that it was becoming quite burdensome for participants. And then we also shifted that to bi-weekly for some period of time. And after that we were contacting our participant monthly. And then we conducted a exit questionnaire which started in early November and finished at the end of December 2020. So the data that I'm going to be presenting today in relation to this analysis I'm talking is only looking at the baseline and the exit questionnaire data. So the two waves of data during the pandemic that I will be talking about. So generally these were the things that the CLSA COVID-19 questionnaire included. These questionnaires are all on our CLSA website. I don't have time to go through each and every question we have. But generally these were the constructs that we focused on. We looked at COVID symptoms. We looked at COVID status, whether they were tested positive. We looked at risk factors. We looked at the healthcare use, health behaviors, what type of public health, how they were sort of engaging or using different public health measures, social factors, depression and anxiety. Just to be clear depression was, anxiety was not asked at, no depression and anxiety were asked both at the baseline and exit questionnaire. They were not asked in between for the weekly and bi-weekly or early monthly questionnaires. We also asked them questions related to economic consequences and in the exit questionnaire that was done in November to December, we also added additional questions related to function and mobility. So as part of this, this particular team of researchers, we worked very closely I mentioned earlier, we started to look at analyzing different aspects of the data that are emerging. And to be honest, we are still trying to understand the data and the nuances of the data. But we did focus on looking at some of the issues related to mental health, specifically the depression. So the reason we focused on mental health or depression in this context was that the, all of this was happening against the issues of pandemic restrictions and people being isolated in their homes or restricted to their own homes, not meeting with their family members or other social connection was happening with the backdrop of existing physical mental health morbidities, social isolation that existed prior to pandemic loneliness issues. And some other people who had caregivers or family or, or, or formal caregivers losing connection with that. So the question was all of this that is happening in, in, in the context of what people already had, and that these new emerging situations that were coming from the pandemic itself, how that is impacting the mental health specifically depression and anxiety in older people in the CLSA. I'm only going to be talking about depression today, anxiety, maybe some other times. But that's what I'm focusing on for this presentation. So we sort of started to look at some of the research questions that we will like to address. Obviously, we'll like to look at much more nuanced research questions that the data that we collected during the pandemic are limited because we didn't want to burden a lot of our participants. Nonetheless, some interesting findings have emerged, nothing unexpected, but still interest, interesting nonetheless. So the first question that we were interested in looking at to examine the relationship between social determinants of health and health related factors with changes in the prevalence of different depression and then we wanted to see are these are these association modified by time and time here I mean the pre pandemic versus during the pandemic. And the second question, which we have been working on for the last little while is looking at how the pandemic related stressors, the data we collected during the pandemic, experienced by the older adults are associated with the severity and trajectory of the depression. So I'm going to be talking about these two main questions today. Just to say we use the, the conceptualization of depression is that we use the 10 item center for epidemiological study short depression scale this is the same scale that we have used in the core CLSA. And many of you probably know each item includes four response categories ranging from zero to three, rarely or never less than one day some of the time one to two days, occasionally three to four days, and all of the time five to seven days. Total score ranges from zero to 30 with higher scores indicating greater number of depressive symptoms for our analysis. We dichotomize based on the recommendation of the scale developers that if you have a score of 10 or more. You are indicated to have your screening for depression and that you more likely most likely you have depression. So this is how for the purposes of this study we have classified. Depression when people have score of 10 or more. Obviously, we did not collect all the data as part of the COVID-19 questionnaire study. Some of the variables that I'm going to be using in my analysis was brought from our baseline and follow up one. Cycles of the core CLSA and these are the core areas that the variables that we looked at age, sex, ethnicity, household income, dwelling type, living area. How many people live in your household social participation, alcohol consumption, smoking status, physical activity, number of chronic condition, loneliness. And in some situations we looked at COVID-19 status and I'll come back to that later later. And to examine the first objective, which was to look at the relationship between social determinants and health related factors. We obviously did some descriptive analysis and I will provide some overarching picture what the data actually looks like. And then we also did a way to generalize estimating equation analysis, which is basically a logistic regression modeling that accounts the longitudinal data. And specifically, and it allows modeling for change in prevalence of depression, depression over time. And it's important to note that this does not look at individual level change, it only looks at population level change in the depression. The reason we used weighted GE that it allows one to handle missing data when it is missing at random in a much easier fashion as long as the missing data are monotonic. What I mean by monotonic is if we had a missing data at baseline of the CSA core, but we had data from individuals from follow-up one and during the COVID, this model can't handle that type of a missing data. For every person we have to have a baseline value present. And the flexibility that the GE framework allows that you can use wide variety of link function. And in this case, we use the logistic link function. And this analysis was performed on 39,604 individuals. And these are the people who had the monotonic pattern of missing data. And then we define time period as a pre COVID 19 period as data from the CLSA baseline and follow up and COVID 19 baseline and COVID 19 exit question. So there are in this particular analysis for three time points that we consider. And age was considered as a time varying courier covariate in the analysis. And then we also looked at interactions that that sort of answers the questions related to the interaction between period and some of the demographic or social and health related factors. And this gives you just a bit of an overarching picture of the data and this is CLSA baseline CLSA follow up CLSA COVID 19 baseline and CLSA COVID 19 exit question. And you can see the sample sizes across the top. And it sort of tells you a little bit about the distribution by age, by sex, by household income and the, and what the number of people who screen positive or negative for depression on the CSD 10 scale. I just to keep in mind in the, in the CLSA COVID 19 baseline or COVID exit question that we did not ask the income household income question. So the numbers you see are from using the data from the follow up one, the 3.84 and 3.80 that is based on the, on the follow up one income data. So first looking at some of the, of the, the longitudinal descriptive analysis, the, this is a change in prevalence of depression by period of the top part top chart and the bottom one is for sex. And here you can see that at the baseline that the, that the positive depression was around 13%. It was roughly the same at follow up one, but then it jumped to almost 20% at the CLSA COVID in 19 baseline and a little bit more, but roughly the same at the COVID 19 exit questionnaire. And if you look at the by sex, that the relative increase in female was much higher than male, but both sexes had an increase in the, in the depression. And if we look at by age group, and here we see that again, you see the CLSA baseline, which is the blue bar orange is the follow up one gray is the baseline COVID and and the green is COVID exit. You see actually the relative increase in, in depression scores or depression screen positive was much, much higher in the younger age groups than it was in the older age groups, but it, it, it happened across all age groups in our data. And this is the one of the interesting slides that even though it's a descriptive it's also holds true when we do a multivariate analysis, changing prevalence in depression by period and household income. If you see the less than 20,000 group of people, they had high depression proportions, even prior to the COVID 19, they went up a little bit, but they were already at the high end of it. So the relative increase there was not that big. However, if you look at the higher side of the income, which is 150,000 or over 100,000, you basically see that the jump was much higher, but still much less than the low income people on the whole. So it affected all sorts of people, but it if it that the pandemic affected the mental health of low income groups substantially because they already started at a higher level. And then we sort of started to look at other social participation and living status. You can see the depression rates were already higher with people with low social participation prior to the pandemic and went up substantially during the pandemic. And it also affected people who had high social participation, but not as much as it affected the low social participation group of people. And here you can see the people who were much more impacted by the pandemic in relation to depression were living alone versus not living alone. And if we look at the marker of health we used here was the number of chronic conditions, the notion of the multi morbidity. Again, substantive impact as the number of chronic conditions that people reported having and this is all coming from the baseline. Those group of people had a higher impact in relation to the mental health issues, but it was again happening across all all groups on this slide. So, after this descriptive analysis we decided to look at the, that the way to GE analysis where we wanted to see some of the adjusted relationships because what I showed you before, but we're not adjusted for anything those are mostly descriptive. And here the first, these are the five stars I picked for the discussion purposes here that you can see that in comparison to the baseline co, the comparison between the covert 19 baseline depression score in relation to the pre covert depression score. The odds were roughly around 1.8 and from coven 19 exit to pre covert 19 was almost two. So basically what it is saying that the people went there, odds of getting depression did go up in comparison to the pre pandemic time period. Over the eight month period, the 10 month period when we were collecting data, it didn't go up. It actually stayed pretty high. If anything, it went up a little bit. It will be interesting. If we are able to collect additional data in the coming months to see what happens with the current pandemic wave that we are seeing to the to the people's mental health. Here again, as I mentioned before, the younger age groups had a higher odds, low income groups. Again, three or more chronic condition twice as likely to report having depression and loneliness, which I didn't show you the descriptive was one of the stronger risk factor for for experiencing symptoms during the pandemic. And then we started to look at some of the interactions that I mentioned earlier on. Here you see that again, if you look at the time period, that is the covert 19 baseline versus the pre covert 19. And we look at the interaction between the time period and the income variable, let's say here, that's the second bar, the chart at the bottom of this slide. You can see the people in this that the, the lowest income group in comparison to the highest income group had odds ratios of four so that it disproportionately impacted people in in low income groups, even though it was also higher in higher income groups, but it had a disproportional impact in the low income groups. And this pattern was seen with all sorts of different. You see that up on the top slide, male versus female. Again, female experience much higher odds of experiencing depressive symptoms. And you see this pattern with multitudes of factors, but not as strong as what we saw in relation to income. Again, low versus high social participation, two and a half to almost 2.8 odds ratio living alone versus not living around to and and some of these things either went up with time or sustain themselves at the same level. Here, whether you are looking living in a household standalone house, or you're living in apartment or some sort of congregate setting type of a living arrangement, your odds ratio for experiencing depression was much higher in comparison to people who were living in detached homes. Again, here you see the differential impact of chronic conditions over time in in relation to the pre pandemic to during pandemic and and as the time went on, the impact of the chronic condition actually start started to be get a little stronger. Here we have odds ratio, I think 4.5 versus four. So that was the data related to the first research question. And, and the second question that we were interested was looking at how the pandemic related stressors experienced by older adults associated with severity and trajectory of the depression. So this is a slightly different take on the same question. So the, what is the, when I'm talking about the, the stressors that people experienced during pandemic. There was a series of questions that we asked in our baseline as well as in the exit questionnaire, related to different forms of stressors that people might have experienced health stressors whether they were themselves ill, someone close to them was ill or somebody died. It was difficult to be the accessing resources lots of income, unable to access necessary supplies of food unable to access usual health care or get prescription medicine medications. Questions about conflict whether there was a verbal or physical conflict separation from family and caregiver responsibility. So we sort of. This was not categorized like this in the questions that we asked, but these are the labels that for our analysis purposes we have given calling them health stressors, difficult accessing resources conflict separation from family and caregiving responsibilities. And, and as the bottom of the slide, we say that these experience were grouped as a yes if the participants endorse at least one experience in this specific category, and if they said no, then they were classified as a that no was not endorsing any category in that group of stressors that we have defined here. And here we actually looked at a slightly different type of analysis. This was a latent class growth modeling. It is a group based modeling strategy, which identifies distinct classes of individuals who follow a similar pattern of depression in this particular analysis. Over the four time periods that is CLSA baseline CLSA. First follow up COVID-19 baseline and COVID-19 exit. These trajectories were modeled using a proct fredge procedure in the in the SAS. And COVID-19 stressor experience of loneliness and all of the covariance, which we wanted to a desk for that we saw in previous analysis were added to the model. Model selection involves a bit of iterative estimation and some qualitative judgment. And but mostly, we use some of these fit fitting model criteria to determine which is a good fitting model and which is the model that we want to pursue. One of the advantages of advantages of this growth modeling. This latent class growth modeling is that when we have multiple groups, depending on what the shape of the curve might look like. You can use a linear term quadratic term or a cubic to fit the best fit fitting model in our case. It was actually came down to linear or a quadratic term term depending on what particular group we were looking at. This is where the latent class, it looks at the data it defined. It tells us that the there are three underlying categories of groups of people that experience depression. And based on the best fitting model, we actually defined this into three categories, the three groups of trajectories. The bottom one is the red one is the low and consistent. They were fairly low, but didn't change much over time. The green one is a moderate and increasing, and then the high is the blue one is high and increasing substantially. Just to remind you, based on the CSD, anyone who has a score of 10 and over is classified as having depressive symptom. And you see the high and increasing all have scores higher than 10. And the dotted lines are 95% confidence intervals around each one of these trajectories. Again, just to keep in mind that this is not defining an individual level trajectory. This analysis also looks at group of people who belong to that particular trajectory, the interpretation of the result is at a group level, not at the individual level. So, focusing on the, these are the all other variables that we adjusted for. And this was quite interesting majority of the health raters stressors did have impact on on individuals and here, what we are doing is the first set of columns I had we are taking the this green line moderate comparing it with the red one, and then comparing the blue one with the red one that's why we're the comparison happen so here after the groups of trajectories have been defined. We basically ended up running a multinomial logistic regression. And here comparing the moderate increasing group to the low increasing group, but these are the odds ratios linked to different categories of stressors. And here, the interesting part is that the conflict stressor was almost four times of in the moderate in comparison to the low, but that that effect size or odd ratio goes to almost eight so substantive impact on the mental health of the people who are actually going through a conflict whether it is in the same household, or outside the household we were not able to differentiate but it shows that there is a substantive impact. And then the again, loneliness issue is a substantive risk factor, especially for people who already had a higher level of depressive symptoms prior to experiencing the pandemic and going from five from the moderate to low to almost 15 other issues of 15 for that group. So this analysis gives you a bit of a different perspective that sort of looks at the severity of the depression and how that depression sort of impacts over time through these trajectories when related to these specific variables. So that's all I had to say today, basically in conclusion, COVID-19 pandemic is at a significant impact on the mental health of older people and I think it's important to note that especially the young old the 55 to 64 year old actually had higher proportion of depressive symptoms. Our results show that the prevalence of depression increase during the pandemic, as compared to the pre pandemic. And the increase and it sustained itself through the eight or month follow up that we did during the pandemic increase in depression is not evenly distributed disproportionately impacts some of the socially disadvantaged or people who have health issues. And the group of individuals who experienced pandemic related stresses such as conflict had much higher odds of being in a group that had the worst trajectory or depression. So, I will stop here and hopefully we can. I wanted to leave some time for question answers if there are some interest from the group. But I gather that there is some CLSA propaganda material here as well. Here is a on the CLSA website. We have a COVID-19 dashboard where you can go and see other variables descriptive data on our COVID questionnaires. And we will be adding additional follow up question. I think this is still based on the baseline COVID, but others are coming. And secondly, there's, we have decided to release the CLSA COVID questionnaire data to research community and some of you have probably applied as part of the current deadline that just recently closed. So they are available to other researchers to access for future research questions. And the CLSA core is funded by CHR and CFI and many of the provinces and universities across the country. Thank you. Thank you for Mender for the excellent presentation. As always, I'd like to now open it up to questions. Just reminder muting will remain on, but you can enter your questions into the chat box in the bottom right corner of the WebEx window. I didn't see any questions yet, but I'm sure there will. Here they come. They're starting. So we have a question from Teodora. What was the number of females and males in the study? And then also maybe you can comment if the difficulties because of language barriers and not knowing the system. So here, this slide shows what the distribution by sex is. So you can see the CLSA core itself have roughly 51% female. And that is similar to that in follow up one in the COVID baseline, we had around 52 and COVID 1953. So that's the male to female ratio in the CLSA and during the COVID questionnaire. What was the other question was please repeat what conflict refers to this was a single question we had in the CLSA. That asked people about about whether during this pandemic period. They experienced an increased verbal or physical conflict. So that's what I labeled that that construct as a conflict. It might not be the right label. That that's what we ended up labeling. So it is based on a single question. That we sort of looking at that data. Do you remember what the question was? There's a question. What, how was the conflict question worded? And when was that question asked? Pardon me. It was asked at the CLSA COVID baseline and then repeated at the CLSA COVID. Exit and the question was the, which of the following have you experienced during the COVID 19 pandemic? And then it asked these experiences that I've listed in on this slide and the question was, have you experienced increased verbal or physical conflict? It doesn't ask inside the family or outside the family. It just simply phrases the question in that fashion. Another question we have is understanding that data analysis continues. What ideas have emerged about how to mitigate depression and anxiety across these age groups? Well, we are barely scratching the data. And so right now we are looking at. I think obviously some of the mitigation strategies that will have to be targeted are going to be how it differentially have impacted young old people in low income groups. What kind of support structures they need in order to manage some of the stressors that they have experienced. I haven't really thought about what would be the interventions or strategy that would be needed to mitigate our goal up to now has been thinking about where are these risk factors and who is being impacted more than others. The people who work in the area of mitigation might have to look at these data and say, these might be the strategies that might happen. Obviously, if we look at the caregiving issues in this particular slide that is on the screen, it has impacted substantially older people because they are not able to see their grandchildren, their children, or their other loved ones. So the question is if we have to go through this type of crisis situation again, I bet we can and should be able to come up with plans that the caregivers, especially the family caregivers become part of the care system of the older people who might be living alone or might be living in the institutional setting. Mind you, in our data we have not looked at the institutional setting because these are all community development people, but nonetheless the issue remains the same. Low income side, I guess people have to deal with multiple issues in relation to not having resources, maybe living in small spaces, maybe living in multi-generational households. There might be multitudes of factors that might be driving their mental health issues in relation to the pandemic. So this is just scratching the surface, there are lots more to be done, whether CLC can answer all those questions remains to be seen. But those are some of my thoughts at this point in time. And I think that touches on one question that we had about what the findings may recommend for agencies working with seniors in the community and also prevention. Anything else related to that that you might. I think what it indicates is that the seniors who are in the low income households require special attention. So there needs to be some targeting done there. Other one is the seniors who were lonely. I think that again emerges as a important factor. We do have to make sure that we community organizations or public health agency of Canada or provincial agency think about that the loneliness that existed prior to the pandemic. How that actually increased the risk of some of the mental health issues and how do we think about providing that support to those people in these very, very difficult times. So I think you can pick multiple areas where story is that people need support and how do you provide the support that supports them through this time becomes critical. It's also, and I know some of our colleagues are looking at analysis. It's also interesting that the mental health issues are much more prominent in younger people. Maybe we need to do some additional analysis and understand the resiliency of older people here that their depression rates didn't go up as much as some of the younger people. And I think there are other things to be explored here. Yeah. Okay, switching gears to some of the, the, the data you presented just a question about the odds ratios and whether they were significant because they didn't see a P values. Well, I will for say I'm not a big fan of P values. That's why we provided confidence interval, but to give you a sense that if you are interested in P values, all of them were significant. Except for one that change after we adjusted for the loneliness and that was the living alone variable that became non significant, but everything else was pretty much associated statistically with the outcome. And what about the impact of the cognitive status of participants. We haven't looked at that. Yeah, that's what I thought. Okay. I'm working for there would be some analysis that we might look at in the future. The, the previous cognitive history, but we didn't collect cognitive data. As part of the COVID questionnaire, but as you know, we are still collecting data as part of the follow up to and there's close to 8,000 or so people or 10,000 people. On whom we have collected questionnaire data during pandemic, including telephone batteries of COVID. So we probably have opportunity once the follow up to data become available to address some of those questions as well. So I'm going to read this next one. It's rather lengthy. Do you know if any of the other international longitudinal studies on aging were set in locations that chose to adopt different public health measures at the outset of the pandemic. They add that it is distressing if not surprising to see the extent to which mental health was eroded and so important moving forward to be better prepared to respond to future disasters. Comparative analysis with other settings could be helpful. Any comment on that. Well, it is, first of all, it's really hard to figure out what strategies, what different jurisdictions or countries used. We couldn't even figure that part out easily in Canada across provinces. It's very hard thing to get data to determine what happened. One of the things that other studies, there are other studies that have come out and have shown the impact on the mental health of older individuals in different countries. But majority of the study, maybe there's one or two other studies that I have looked at have looked at or collected data just during the pandemic. I didn't have the data related to the pre-pandemic period that we have the advantage in relation to the CLSA. So the comparative analysis between pre-pandemic and during pandemic gives a special, I guess, interpretation of the CLSA data. But the data that have been collected longitudinally over time during pandemic are showing pretty much the same pattern in different parts of the world. So it is not just that there is something in Canada that is happening. This is the same pattern people are seeing in other countries as well. Even in countries like Australia, there's a paper that has come out of Australia which did not have as stringent public health measures for a longer period of time as we did here in other countries. But during the lockdown, they experienced the same, same impact. I haven't seen data during the opening, after the opening up of the restrictions, but the earlier on the patterns were the same. Another question here, you have a higher proportion of over 75 people for COVID baseline and exit compared with earlier waves. Could that contribute to higher percent of depression for COVID baseline and exit? So there's two things happening and that's a good point that's the question that remember we are picking people from the core CLSA and CLSA is aging. So we have taken that into account so that the higher proportion of people in older age groups as a function of the design of the CLSA. That's what I was trying to say. The age increase, we actually adjusted for age as a time varying core covariate when we are looking at some of the relationship with loneliness or income and other factors that is adjusted for age as a time covariate factor. So I personally don't think that is the reason that we would have seen higher proportion. First of all, I don't think we see a higher proportion of depression in the oldest age group. We see it in the youngest. But I whatever we see in that age group is not just function because we have more people in the 75 age group. At least that's how I am answering that question right now. I will have to think about it a bit more. We have another question about separation from families and that she says, I'm surprised that separation from families did not figure more prominently into your results during COVID lockdown. Any thoughts as to why? Well, we have this is the analysis I presented and separation for family showed a fairly moderate impact on the trajectories of worsening mental health. Do we have a much more elaborate data in the CLSA? No, simple question that people responded. I think some of these questions might be better answered with the, when we have CLSA follow up to data. As the questions are about the families, families, separation or caregiving a bit more nuanced in our data collection than they are in this one. But even with this single question, at least in the community developing at least the participants that are contributing to this data had increased odds ratio, but not as high as let's say conflict or loneliness. I just wanted to say if I remove the loneliness variable from the map model, the separation from family does go a little higher in its effect sizes. So loneliness probably explains the majority of it. What about data on perceived stress levels? Do you have data about perceived stress levels? I'm trying to remember. I can't. I don't think so. I don't think we collected that as part of the COVID questionnaires. But I don't think it was part of COVID, but maybe main wave. No, we don't have a perceived stress in the main ways we have. We can create a perceived stress variable in the previous from the previous waves, but there is no single questionnaire that has the perceived stress per se. We're getting towards the end of the hour and there's just a few questions left. If I just wanted to note, if we didn't, if we don't get to your question. We can follow up with you via email with some further comments about it. A question from Larry Chambers, which I'm not surprised he's asking this question. What actions are, what actions of government are informed by these results? If any. And we have there been any work to charge on at this question already earlier on. I think what this data identifies that the mental health of older people or the people over the age of 50 people over the age of 50 in the CLSA is not affecting people equally. There is a population heterogeneity. There is a different social determinants of health that the different social factors impact people differently. So I think these data probably not necessarily identify anything new, but reinforced that the pandemic we know from the infection point of view it has impacted population very differentially. And the same thing is happening in the context of the issue like mental health. It affects people very differently. People might be writing the same storm, but they're all in different boats in context of mental health. And I think that's what these data be a firm to some extent it identifies some of the, some of the new things for example that which to me was surprising that even after the first lockdown when we had the easing of the lockdown, the depression, the proportions of people with depression remain pretty high. It didn't go down. Question is how long it will take before people start to feel back to their original mental health or is it something they're going to have this for a long period of time we have some PTSD measures in the CLSA that will be interesting to look in the future how it has what impact traumatic impact it has on people's lives but I think that these data at least to me it emphasizes that the lockdown or public health measure which are important to have to prevent infection has had a differential impact on different groups of people and the mitigation strategies have to be targeted at those groups of people and mental health issues depression issues are are important. We are doing some other analyses in relation to persistence of symptoms. You see that game same population heterogeneity and differential impact on on on groups of people from low income males to females, people living alone versus not living alone. So it is a pattern. That is the social disadvantage groups are really have a experienced cost of them heavily, not only from infection but from other consequences the pandemic including mental health so maybe some of the strategies have to target those individuals. Okay, well, it's 101. I just want to also remind everyone before people start to leave if you can please complete the poll before you leave. I think a good way to end parminder is two last questions one that was asked initially. I do. I'm sorry. I have a call with the minister so I do have to go. When will the results be published. That's what I was going to ask. Well, the first paper the that the gene analysis is under review right now we'll see if anybody accepts it and the other second objective is being written up. Okay, great. Well, thank you very much and to everyone who we didn't answer your questions we will follow up with you. A few closing remarks before people start to leave. I'd like to remind everyone that CLSA data access request applications are ongoing. The next deadline is September 8th. If you can visit our website under data access to review the available data in including any of the COVID-19 questionnaire study data as well as additional details about the application process. I already reminded you to complete your survey. The upcoming webinar for May is entitled a long term cognitive impairment following concussion findings from the CLSA. It will be presented by Dr. Mark Bedard who recently defended his PhD thesis in clinical psychology at the University of Ottawa. You can visit our website to register for that if you're interested. And remember the CLSA promotes this webinar series using the hashtag CLSA webinar and we invite you to follow us on Twitter at at CLSA underscore ELCV. Thank you again for attending our presentation today. I know we're a couple minutes over. But we look forward to seeing you and Dr. Bedard at the next webinar. Thanks everyone.