 and a statistical analyst in the aging community and health research unit. Her role involves active participation in shaping the undergraduate nursing curriculum as it relates to health sciences and evidence-informed decision-making. Her goal is to ensure current content, scientific integrity, accuracy and rigor in all courses within the health science portfolio. Thank you very much for being here and I will go ahead and give you control of the webinar and invite you to begin. Thanks Carol and thank you for those of you who are attending today. What I'm going to share with you are the results of the study we've done using CLSA data and looking at the relationship between multimorbidity and disability and particularly the role of mental health in that association and trying to understand and unpack the potential contribution of mental health associations. First of all, I'd just like to acknowledge the research team, so from McMaster in addition to myself, Lauren Griffith, David Cantors, Montana Fisher-Shotten, Maureen Markle Reid and Jenny Plough, as well as Andrea Grunier from the University of Alberta. Today what I'd like to share with you, first of all, the purpose of the study, a little bit of background on what we know about different areas, so multimorbidity, disability and mental health, not necessarily the intersect of these different areas because there hasn't actually been a lot of work on that, that's subject to this study, but we know some things about each of these areas that sort of put our radar up when we're doing our work as to what we might need to control for or look for associations with, what might confound relationships that we're looking at and so on. So I'll share a bit of that and then the objectives of the study, some of the main hypotheses, we had a number of them but the main ones in this work are results, discussion and what we're able to conclude from what we've done. So the purpose of the study was to improve our understanding of really the differential impact on disability of mental health conditions, coexisting with other physical chronic conditions and so maybe to better understand that it's really a multimorbidity study, so it contributes to that literature, but we're focused on mental health in this work particularly. So for those of you who aren't actually familiar with the term multimorbidity, it's defined as having two or more chronic conditions, so we're not looking at acute conditions and sometimes it's even defined as three or more, but I think two or more is the more common definition, it's basically having multiple chronic conditions and however you define multiple and then broadly we're talking about physical chronic conditions like COPD, cancer, coronary heart failure, arthritis, diabetes or mental health conditions, depression, anxiety, bipolar, schizophrenia for example. So those would all be collected as your multimorbidity and then some background on the research in multimorbidity. So there has been a fair bit of work done actually at this point. It is recognized now as a global health burden, some people actually call it a chronic condition in and of itself, but it's associated with a number of negative health and other outcomes. So disability, more disability is associated with higher levels of multimorbidity, more mortality, more complexity in terms of clinical management or self-management and higher healthcare service use and cost. So there's been research showing those associations with morbidity. We are focused here today on disability. There's also been some work showing that mental health disorders themselves are linked with higher levels of disability in comparison to physical health conditions and then having physical health multimorbidity itself is linked to having a higher level of mental health disorders. So multimorbidity goes together and physical conditions, when you have them, then you're more likely to have a mental health disorder. There was just some background noise there. We weren't sure what that was. With respect to mental health now and disability, there's been some sort of, I would call it synergistic impact of mental health research that are findings that have occurred and they're all coming, not all, but in QAnon's is one author, Anna Quinans. She's from Europe, I believe, but she has done a lot of work in multimorbidity and disability. So some of her work is in particular pointing to a sort of synergistic, I guess, impact of mental health. So for example, she has found that disease and her research team, disease clusters that include depression are associated with higher disability compared to disease clusters that include only physical health conditions. She also did some work where she looked at the most prevalent disease clusters and limited that to the top 14 clusters of which one and only one of them actually included a mental health condition. It was depressive symptoms and depressive symptoms combined with hypertension and arthritis, although they were ranked 11th out of 14th in terms of prevalence. So not the most prevalent condition, it actually was associated with the highest level of disability. So there's that finding out there. Again, it's sort of suggesting there's something about mental health that elevates disability relative to other conditions. Then there's also been some work where people looked at index conditions. These are one of studies, arthritis, COPD, diabetes, where again the finding has been that when those index conditions are combined with a mental health condition, typically it's depression, that the disability is higher for those combinations of conditions than those index conditions combined with either physical, other physical conditions or on their own. So again, when suggesting there's something unique about mental health, and then also some things we faced in this study and multimorbidity research studies in general face, there's a lot of other things, associations and factors that we need to take into account. So studies on multimorbidity, disability and mental health have also shown associations will associate demographic factors. So multimorbidity and disability, they both increase the age and have shown that repeatedly. Multimorbidity and disability are higher in women, are higher in lower SCS or socioeconomic groups, minority populations, mental health conditions are higher in women. So we need to pay attention to some of the associated demographic factors because they may be confounding associations that we're looking for. Excuse me, can you just ask people on the phone listening to mute their phone while they're, well, we're doing the presentation. Thank you. Okay. And so another application with multimorbidity research is the lists that are used to identify whether people actually have multimorbidity. So they vary widely in terms of the conditions that are actually included on that list, how many and what types. And then in particular mental health conditions are quite often not included on those lists at all because it's difficult to actually get information on mental health conditions. So in administrative data or self-report for a variety of reasons. So that information often isn't available and so mental health hasn't been within multimorbidity studies looked at very much. And then finally disease clusters. We had intended in this work to do a fair bit more on clusters of diseases, but it is very hard to do. And there are a number of things, especially when you have a large list of chronic conditions and we did have a large list. So that sort of exponentially increases the clusters of conditions that you have and look at. So you have to limit those, well there's a few problems. You have to limit the number of conditions you look at effectively and so you start looking at the most prevalent ones or just dyads like a pairing of conditions and stop there. And then you also are running into issues with trying to associate people or link people with clusters uniquely and that requires a lot of assumptions that aren't necessarily going to hold. So it's been hard to actually do disease clustering research fully. Usually we pick and choose the clusters we look at. So it's a partial picture of association that we're actually seeing. So now to this study what we were really doing in terms of our specific objectives we want, first of all, wanted to determine the rate of physical disability and mental disorders that were in the CLSA baseline data set and then also to see what mental health conditions clustered with in terms of what other physical conditions they were clustering with. Then we wanted to look at the association between multi-morbidity and disability and really try to isolate the role of mental health in that association. And then finally to try to investigate age, sex and some of the other socio-demographic factors that we know are also involved in the relationships between disability and multi-morbidity and what we should be controlling for and what associations might be modified by those factors. And in terms of hypotheses we had quite a few but the main ones in this work there was one relating to disability and then one to the clustering. So in terms of disability this first hypothesis was that multi-morbidity combinations that included mental health conditions would be associated with higher levels of disability compared to combinations of disease clusters that included only physical conditions. And we actually were looking at or thinking this would be true for each level of multi-morbidity. So leveling of multi-morbidity in our case was measured by count of chronic conditions. So what we were thinking is if we looked at each level or each count within the count we would still see a differential impact of mental health conditions on disability versus people that don't have mental health conditions. So that was one hypothesis. And then in terms of clustering what we were expecting to see is that mental health conditions would cluster with more symptomatic conditions like arthritis, COPD, stomach bowel disorders and that might help us to understand the link with disability. And there's been some research that has actually shown these connections so when mental health is clustered it has clustered with particularly arthritis and COPD for example. So we were expecting to see that here as well. In terms of the data that was used in the study we were looking at baseline CLSA data and there were 51,000, slightly over 51,000 people in the cohort. These were communities well in Canadians aged 45 to 85. And the data were obtained through either in-person or computer system telephone interviews. So for those of you familiar with the CLSA data this data set included both the tracking and comprehensive cohort. Measures that were used in the study so our outcome was disability and that was a dichotomous measure so if a person reported any limitation and there were 14 either basic or instrumental ADLs that activities are daily living and if they reported any of those 14 then they had a disability versus didn't. If they didn't self-report any of those limitations. And these items that were the actual limitations came from the ORS instrument for those of you that are familiar with that instrument. The multi-morbidity as I mentioned was captured by a number of chronic conditions. So people and chronic here was that you had a condition for at least six months and then participants were asked as a doctor ever told you that you have and were asked for a number of different chronic conditions which I'll mention in a moment. And mental health specifically here was mood and anxiety self-reported among the chronic conditions that I mentioned in point 2 about so mood and anxiety we pulled out and that was how we measured having a mental health condition. We also though did some work we added to that so in addition to self-reported mood or anxiety we actually captured depressive symptoms and because the CESD instrument was also captured by the CLSA data so those people that poured 10 or more on the CESD instrument were also in another analysis added to the cohort of people with self-reported mental health conditions. And so we listed mental health with that broader composite definition. And the reason for doing that is again mental health is often under-reported or under-diagnosed. So we thought we would possibly be capturing more people with depressive symptoms included as a measure of mental health but also try to capture people who more than likely would have depression but may not even know it. The doctor hasn't picked it up either. So we did that composite work as well and just so you can have a look at all of the conditions so we had quite a large number of chronic conditions that were used in measuring multimorbidity for this data set and this is showing you the prevalence for the 51,000 and some odd people from high to low and again even this list is actually a grouping of other, of more conditions that were captured. So for example arthritis is both osteo and rheumatoid arthritis captured together here. Thyroid is hyper high pole so heart condition, there were a number of heart conditions grouped together under here. So this is actually itself a grouping of chronic conditions. And methods, what we actually did, so we actually did this work in phases. So phase one was looking at disability prevalence for people with and without mental health conditions and what we were doing is actually comparing people at the same multimorbidity level, so at the same number of chronic conditions level, those with and without a mental health condition and looking at whether there was a difference in disability prevalence between those two groups holding multimorbidity constant and then we looked at some stratified analyses by age and sex. We also used logistic regression to estimate the odds of disability for people with and without a mental health condition. So the independent variable there was a number of chronic conditions and with the reference category being zero chronic conditions and that was true in all the models, a dependent variable with disability, yes you have it or no you don't. And running these models separately for people that had mental health as part of their multimorbidity versus didn't. And then again those models we first ran unadjusted and then adjusted for socio-demographic factors that looked like they should be at least controlled for in terms of the association between multimorbidity and disability. And then in phase two we captured more mental health conditions by also including using this composite measure of depressive symptoms plus self-reported depression or self-reported, really that should say mood or anxiety there. And then disability, we also looked at disability the main analysis, the phase one analysis, disability with IADL and ADL either. But then in phase two we actually separated them out because we just weren't sure if we would see a difference. IADLs are more cognitive and they're more normally more highly prevalent and often are emerged earlier than ADLs. So there are these differences and we weren't sure what that meant for our analysis so we actually looked at them separately. And then phase three was our factor analysis to look at multimorbidity clusters and what physical conditions mental health was clustering with. And there was also this one variable that we thought well we could actually see if there was a mental health connection. And that was a social participation variable. So CLSA captures a variable that asks whether you had a health condition that restricted social participation. So we could actually stratify that by those with and without mental health conditions to see whether mental health might be affecting social participation. So we looked at that too. So our first set of results for phase one was 51% of the cohort were female and 20% had a mood or anxiety disorder. We also saw that overall multimorbidity was so people had 2.2 chronic conditions of which the actual average was higher if you had a mood or anxiety disorder than if you didn't. So we are seeing in the literature that if you have a physical condition you are also more likely to have a mood or anxiety disorder as well. And then just here we just looked at some of the main demographic variables and we stratified. So we looked at all participants, all 51,000 and then stratified did a stratified analysis. Those with mood or anxiety that's the center column versus those without mood or anxiety. The column on the right. And here again you do start to see some of the differential influences mental health right off the bat. So you see sex differences. So women overall in the cohort were 51% but when we look at 51% of the cohort when we look at those with mood or anxiety versus without women are 63% of those with mood or anxiety and 48% without. So more women than men have mood or anxiety disorders with respect to age. What we see there is that the younger age groups, 45 to 54, 55 to 64, more of those have mood or anxiety relative to the overall cohort. And so mood or anxiety are more common in the younger age groups relative to the older age groups. With respect to the other physical chronic conditions the most common ones are shown here. Most prevalent ones are arthritis, eye conditions, hypertension or high blood pressure, diabetes and respiratory. For people that have had any of those mood or anxiety is more common for all of those conditions. So you're more likely to have a mood or anxiety disorder if you also have these physical chronic conditions. So we saw that in terms of number of chronic conditions you can see here that at the higher level 3, 4 and 5 plus more people have a mood or anxiety disorder among that group than at the lower number of chronic condition levels. And then social participation. So that was that variable I mentioned earlier. Was there a health restriction that limited you in terms of your social participation? So overall 7.6% of the cohort reported that there was a health restriction that limited participation, but within those with the mood or anxiety 14% were reporting a limitation as verses 5.9% for those that don't have mood or anxiety. And with limitations the last row there 10% overall reported having a limitation, ADL or IADL, but within the mood or anxiety group that was 17% as verses 8%. So there you're seeing right off the bat some differences among those with and without mood or anxiety. And here now we're looking at disability prevalence. So that center figure figure one is actually everyone combined with the y-axis being the percentage or prevalence of disability and the x-axis being number of chronic conditions. So you certainly see that as chronic conditions increase from 1, 2, 3 up to 5 plus that disability prevalence increases overall. And then when you look at the blue versus gray you're comparing the mood to the not mood. People with a mood disorder, mood or anxiety versus not at each of those levels. So there seems to be when you look overall for some of the levels there seems to be a bit of a difference, a higher prevalence of disability when you have a mood disorder versus don't. So for example if you look at level 2 level 4 level 5 plus you see that the blue bar is higher than the gray which suggests that there is a disability associated with a higher prevalence for those levels. And then when you break it down into women and men, so the figure is on the right, a few things you can see from that. So that differentiation between mood and not mood for men, the bottom figure doesn't seem to be there until you get to the 5 plus. It's pretty well the same prevalence at each level of multimorbidity whether you have a mood or disorder or not. But with women you're seeing more of a separation there, higher disability with those with mood versus don't. And higher disability in women compared to men really at each level of multimorbidity. So you're seeing all of that. So there are sex differences here in relation to disability, in relation to mood and they're coming out in that figure or those figures. And then also we know that age is a factor. So it's not just accounting for sex but also looking at age differences. So here you've got age and sex captured and looking at disability prevalence and what you see are again higher levels of disability in women than men at each of the age groups and quite a difference between the, in terms of disability prevalence for those with mood versus without mood disorders at each of the age levels and for women and men. But we also know this and that number of chronic conditions and age are highly correlated with one another. So as you, so this is just showing you that but so there's both age and number of chronic conditions happening here in this association and we really need to teach, need to capture or separate both, stratified by both to really better understand what's going on. So this is just showing that number of chronic conditions at each of the age groups very clearly is associated with age. And yes, you're seeing that mood versus not mood distinction with number of chronic conditions being higher for those with mood versus not mood. But so this is really just now thinking okay age and multimorbidity we really would like to control for both to understand what's going on and that's what we did here. So we actually stratified by both or captured age separately and then age groups separately and then number of chronic conditions within and looking at the prevalence of disability. So you're seeing generally here that there's an increase in the level of multimorbidity with or the increase in the level of disability with multimorbidity and especially when you get to the older age groups so there's a gradient so with the 65 to 74, 75 to 85 in terms of going from 0 to 5 plus in both the 65 to 74 and 75 to 85. So disabilities increasing with number of chronic conditions in both of those age groups it's a little less of a gradient in the younger group 45 to 54 especially if not as much of a gradient 55 to 64 somewhat of one but it's really distinct when you get to the older ages. And then there's some evidence here that disability is increasing with age so if you actually for a given level of multimorbidity so if you actually compare the bars it's hard to see I think in this figure but if you compare 0, 1, 2 chronic conditions and you then compare those bars across each of the age groups there is a slight increase in them over the ages. So we can see we have a bit of a window into what's going on and then with respect to disability prevalence and so where we went from there is actually to run the logistic regression model to actually capture the odds of disability and again to hear these odds that we're showing you are compared to people with 0 chronic conditions and we're comparing those with a mood or anxiety disorder to those without at each level of multimorbidity and so here you can see that it does appear to be a slightly higher disability odds among those with a mood or anxiety so comparing the blue dots to the dark blue dots light blue to dark blue all the way along they appear higher the light blue dots those of the odds of disability among those with a mood or anxiety disorder so you are seeing that here the problem though being that age is concerning some of what we're seeing we know that because younger people have mood and disorder but they have less disability so we know that it's important to actually control for age so we did that's an unadjusted model you're looking at here here's the logistic regression results when we actually control for age so it's included as a covariate in the model and once we do that we can see a more significant difference between the light blue and the dark blue dots at each level of or number of chronic conditions levels of multimorbidity there's some overlap in the confidence intervals at some of the levels of multimorbidity for example levels 1, 3 and 4 there's some overlap there's actually no overlap with level 2 or 5 plus so that's when we just adjust for age and I'm not showing you all these results because this would go on for a long time but we did quite a few stratified analyses looking at other socio-demographic variables and whether we were seeing any difference in the association between multimorbidity and mood versus not mood and disability for other factors so age and sex we've looked at but we also looked at income and education living alone, social support and I think that David pretty much those ones and so we did look at all of this and we what that all suggested was age, sex and education at the end of the day seems to show the strongest relationships with disability so we wanted to be sure in our work from this point on that we at least adjusted for those covariates in the model so here is that the results for when we do actually adjust for age, sex and education and again it's not really changing the overall result that we had seen which is slightly higher disability levels for those with mood versus without at each of the levels of multimorbidity so that looks a lot like what we were looking at when we adjusted for age. Now getting into phase 2 so this is now considering other people so the under reporting of mood and anxiety disorder so considering the composite measure of CESD those people that scored 10 or more on that instrument in addition to those that reported self-reported mood or anxiety and we re-ran the models with that considered to be mental health, having mental health conditions and so what you see here is now a clear separation actually between those disability levels for those with a mood versus without at each level of number of chronic conditions and when you actually look at the OR and the confidence intervals what you see is the ORs are higher here than they were for phase 1 for the mental both with mental health higher at all levels of multimorbidity and the confidence intervals now do not overlap at all for any level so that's what happened when we actually broadened the mental health to include stress symptoms. We're seeing more clearly here, higher disability levels with those that have mood or anxiety to find that way and that's actually a model that's adjusted for age, sex and education as well. Those results. And then the other thing we did in phase 2 was actually look at limitations separately so ADL and IADL examine separately. So here are the results for ADL only and so there's a lot like the phase 1 results actually so very much like it and in fact the only real difference is the confidence intervals are tighter here and then they were when you compare this to phase 1 results so it looks a lot like phase 1 with tighter confidence intervals around both those with mood and without and that's as opposed to this one which is looking at those only and an adjusted model like before but here what you're seeing is more overlap in the confidence intervals, more variance actually especially with those with the mood or anxiety disorder. Well right so that's why you have wider confidence intervals as fewer people have IADL. That's what the analyst is saying David who's here too and I would expect that so that's the reason for the wider confidence intervals. Okay and then phase 3 was actually looking at disease clustering so here the interest was in just what are the multi morbidity clusters but then where does mental health land, what does it cluster with and in the research to date it sometimes is clustered with some of the conditions I mentioned earlier arthritis and COPD but there are other studies that actually show that it's on its own completely so we weren't sure what this is going to look like for our data set but here are the results of clustering so this is the rotated factor analysis results which came out suggesting there were four factors or four disease clusters. The one that's familiar and it comes out in almost every study is the cardiometabolic one which includes your hypertension, diabetes, heart condition, stroke our second factor was where mood actually ended up with anxiety disorders which were clustering with bowel disorders, migraine, intestinal or stomach ulcers and respiratory conditions and then there was a miscellaneous one arthritis ended up in actually that factor or that disease cluster with osteoporosis sorry and then finally there was a neurological factor which was UTI combined with other neurological conditions was the last factor so here we actually see mood combining with some of the things we expected, bowel, migraine, stomach disorders respiratory conditions, some of the literature is pointed to mood or anxiety combining with those. We expected arthritis in that group but didn't see that and UTI we would have expected to however what we can notice about those two conditions if we look at the middle column is the factor loadings were actually higher for arthritis and UTI with respect to that second condition where mood fell as versus any of the other conditions in that column so there is some association there of those other expected conditions with that second factor where mood fell and then finally looking at social participation that other variable we had in weather health conditions were limiting your social participation we did look at that as giving us some sort of signal about whether mental health would actually impact your social participation and what we see here are the results of looking at that just briefly so you're seeing that for both men and women there is a difference between those reporting social participation restriction for those that do have a mood or anxiety disorder versus don't across all age groups and both sexes so we are seeing again this differential impact of mood and anxiety with respect to reporting a social participation restriction and here we could benefit I think a bit by unpacking this a little bit more and controlling for a number of chronic conditions but I don't think it would change this main association that we're seeing that mood or anxiety is differentially impacting or associated with higher social participation restriction so to sort of wrap up what have we seen and found in this study a few things so mood or anxiety disorders are more prevalent in younger age groups and in women they cluster mood or anxiety disorders appear to cluster with highly symptomatic physical conditions respiratory, migraine, and bowel on intestinal in this data set where what we actually saw it cluster with disability is higher in women than men at all levels of multi morbidity the prevalence of disability increases with the level of multi morbidity and at a given level multi morbidity so number of chronic conditions the prevalence of disability is higher among those with mental health disorders compared to those without and that is true even when we adjusted for the socio demographic variables that also might have been confounding that association for men and women in all age groups those with mental health disorders were also more likely to report social participation restrictions due to health compared to those without then the one thing we just wanted to point out we were actually expecting I think more of a difference in terms of disability prevalence for those with mental health conditions versus without and we've actually seen some work recent work at some of the conferences we've been to where it does seem to be more of a difference than we actually saw in this work and we're trying to understand that a little bit better and one thing we have noticed that is in some of the work we've seen at conferences is actually the comparison when you're comparing those with a mood or anxiety disorder to those without it isn't always controlled for the number of level number of chronic conditions so quite often the work doesn't control for that so what they'll do is they'll take number of physical chronic conditions and then add a mental health condition to that and then notice that the disability or health services has increased when you add a mental health condition but the problem with doing it that way is you've also added a condition in and of itself and we know that there's a very clear relationship between many outcomes health services disability and just strictly number of chronic conditions so that's why in this work we've controlled for number of chronic conditions but actually we didn't start out that way and this figure is actually showing you where we started which was sort of the same place everyone else started so for example when you look at this figure you see that those with mood this is our original bar chart but you see that those in the light blue are those with mood or anxiety and compared to the dark blue and what we did here is the zero for example was the dark blue with no chronic conditions and the light blue was having a mood or anxiety disorder and where it says one that was one physical condition versus a physical condition with a mood or anxiety so really the one is actually comparing two to one and three is comparing three to two and so on so you're not actually holding constant the number of chronic conditions and so that's why we here we saw quite a significant difference between the prevalence of disability for those with and without lids but that's when after we thought about this a little bit more we started controlling for number of chronic conditions and that then brought things down a bit in terms of the difference. Some limitations that we should acknowledge here are that relatively few CLA participants actually reported ADL and IADL limitations. This is a fairly healthy group it also starts at 45 so we're not looking at older adults and we didn't restrict the analysis to that but they're community dwelling relatively healthy population at this point so we're not seeing we only had 10% of this population reporting in ADL or IADL and the word is not cross-sectional we're looking at baseline so we can't make causality claims from what we're seeing here and also we ran into especially when we started stratifying for sociodemographic factors and a number of chronic conditions and mental health not mental health we're stratifying finding many different ways and creating small cell sizes and that can become difficult so we had to sort of rein ourselves in there in terms of how many stratifications we did because the cell sizes were getting very small and then multimorbidity itself we're actually measuring it as a count of chronic conditions but we know that that needs some unpacking so we know that if you had two chronic conditions for example we wouldn't expect your disability level to be the same if it was anxiety and COPD versus anxiety and hypertension so you'd have two either way chronic conditions but we would expect that disability might be different for those scenarios so we're not getting into the type of physical conditions and getting specific about those combinations but that's there too and so there's more work needs to be done to actually understand I think mental health in relation to specific conditions that you are pairing with so in terms of what we are able to conclude from what we've done these results suggest that the presence of mental health disorders increases the level of disability and decreases social activity at all levels of multimorbidity measured as a count with potentially stronger effects in women compared so not sure if we have questions Thank you very much for a really great presentation I'll open it up now to questions as a reminder muting remains on but you enter your questions into the chat window at the bottom right corner of the WebEx window and I'll read it out and we'll go ahead and have a discussion so we do have a first question from would you comment more on the inclusion criteria for the chronic conditions did you apply the most prevalence conditions I've seen high blood pressure 38% to neurologic conditions 2% is there any reason why you keep more than 2% scroll down here you can read you can read the question for yourself any more comments on the inclusion criteria for the chronic conditions so bending the chronic conditions so we accepted all chronic conditions we didn't actually exclude is that all we excluded okay so the analyst is here saying we took out anything less than 1% so while there are some conditions that are missing here and we did take multiple scrolls and again that was one that very few people reported so we did take out those but otherwise left in the rest so if you have a follow-up question on top of that feel free to go ahead and type it in we'll try to keep that conversation going I have Dr. Lauren Griffith from HCI Dr. Griffith here in the room with us and she has a couple of questions you say that different types of mental health conditions are less often considered in studies of multi-bormidity how might focusing on these two conditions impact our results so as opposed to schizophrenia and bipolar is heavier yeah only focusing on only yeah the reason for looking at mood and anxiety here were they were the ones captured most frequently in studies where mental health has been used in multi-bormidity so that's in part our focus they're also probably much more frequent than bipolar and schizophrenia so I think you'd run into small sample sizes there are small cell sizes when you started including those conditions too and I don't know what they look like I was looking at some of the literature more recently reviewing it again and I've even noticed people will take depressive symptoms but they won't take some of the other conditions like bipolar so it's really been a major focus on mood and anxiety as opposed to some of these other mental health conditions that are less frequent and I think more extreme in many cases that extremity might argue that those would have big effect they might in a small group of people they might so yeah so should we I think the first thing we'd have to look at is whether you know how many people actually reported those and we were trying here to create results that we could compare with the broader literature which would be then the focus on mood and anxiety I would say we have a question from Maya Lynn Lee Daigle why do you think it was more common for mood disorders in younger age groups how can this influence how we focus our efforts and care for these younger age groups that's a great question that is a good question I would imagine the younger age groups younger here meaning the 45 to 64 are probably the working population they have parents that are elderly and taking care of those people as well so they're probably caregivers and so they're stressed in that regard and I think it's likely reflecting that you do see in healthy populations that relationships of lower depression levels with age so that's not uncommon in healthy populations when you look at chronically ill ones those people that have other conditions like diabetes or COPD that's not the case but I think in a general population like we're looking at here actually it's a lower prevalence for the older age groups and that's consistent with the broader literature but then again that still leaves the what can we do about this or what should we be doing about this yeah I don't know David do you have a comment on that if I fit into a question that I had as well which is did you look at the differences between teasing out anxiety versus depression no we didn't and that might change with age quite a bit and make recommendations for each of those age groups different because of connection I think again it's sizes when you only have 20% overall and then you start separating into those with mood and anxiety and pulling them out you might run into small sample sizes again ABL and IEDL having said that ran into the same problem so we might that might be worth separating and looking at that I think I can see that and like I say small numbers you run into the many times throughout these stratified analyses that would be another example but it might pull out a difference particularly for the age maybe Monica's question I'm always curious about the difference in women versus men mental health conditions do you feel there is an element of lack of reporting by men thereby spewing the results and that I have a question there too that you know the change in your model when you include depressive symptoms the C, E, S, D you could see that that affect on men becoming more like women right because of the self reporting it could very well reflect that I mean that's the perennial question in every study you know you are not every but almost many many studies women have higher mental health conditions than men and they go to the doctor more frequently than men normally are higher uses of health services than men so that may all go together as to why women report more than men and maybe the CESD that's what I was thinking originally was the CESD maybe adjusting for that I don't think it's I do think it's under-reporting it's probably under-reported in both groups but perhaps more so in men a couple more questions we have time for Connie Morano asks would arthritis count only once for multiple sites of arthritic pain knees, hands, elbows, back I ask because of the compound effect of pain my add-on question is can you review the cluster analysis that did not put an arthritis for the symptomatic? Arthritis would have counted once for multiple sites because we did have information on arthritis separately by location but we would have counted it once here and same with heart conditions there's more than one thing in there and I wonder though I think maybe what I'm thinking when I'm seeing that question is if we were to separate them out would we see a connection? Count them as more than one? Yeah would we then see a connection with movement so I'd be like we were expecting to see. I think that's probably a valid point One last question here would people with mental health issues less likely to be able to participate in the CLSA possibly causing a survivor cohort effect leading to some of the age associations that we're looking at? That's a better question for our CLSA people here from participation rates did you see that? What we have done is we have compared the CLSA data on number of chronic conditions with the data that we have from other staff cancer base or from the census and we found fairly similar results with many of the chronic conditions where the CLSA would potentially differ not necessarily on depression but certainly on cognition because people had to be able to provide data on their own. We did not have proxies for CLSA at baseline and they had to be able to sign their own written consent so there were some differences but generally when we compared them our population to the general population of Canada in that age range that met our inclusion criteria was fairly similar. Again that was Dr. Lauren Griffiths here in the room. That included general mental health as well as depression I believe was in that analysis. One last kind of overall question. Do you have any recommendations based on your work for researchers studying multi-morbidity future directions? Go after mental health. We do and it's a real focal point for a lot of our intervention studies. I've participated in a lot of intervention studies in our research unit and we study diabetes and we study stroke but mental health is a really significant element of our intervention even though there are people with stroke and people with diabetes. I think it's capturing mental health and if you've got available CESD or instruments like that to actually be able to tap into feelings and perceptions and emotions that don't get diagnosed it's a really important thing to be able to capture so some of this work the work by Garen for example he actually chose not to use self-reported depression instead ran an instrument and captured depressed symptoms and that was the mental health measure. So I think some people are trying to do that because they recognize that self-report is tough to get an act and even admin data it's tough to get mental health conditions. Well thank you very much for being here and I'd like to thank you again for giving us such a great presentation we appreciate your participation in the CLSA webinar series. Thank you very much thanks for having me. So if we can advance the slide and remind everyone that CLSA data access request applications are ongoing the next deadline for applications is June 5th 2019 please visit the CLSA website under data access to review available data further information and details about the application process. I'd also like to remind everyone to complete their survey located under the polling option if you have any questions or concerns that we can help you with write us in the chat box and we can help you out. Thank you for filling the poll we appreciate you helping us focus our future webinar series. And reminder the CLSA promotes this webinar series using the hashtag the CLSA webinar and we invite you to follow us on Twitter. Please go to our CLSA website to register for our next webinar series presentation coming up soon. And this slide is showing our next month's CLSA webinar the 2019 update on the Canadian Montreal study on aging what's new and what's next for Canada's research national research platform on health and aging. Well, we'll have Dr. Parminder Aina the lead PAI of the CLSA. And Dr. Yajonet scientific director of CIHR Institute of Aging here talking about the CLSA platform so that should be very exciting. Thank you again everyone for attending today's presentation.