 I'm here at the University of Texas at San Antonio in our Department of Demography. And we have what looks to be a great session here. There's two papers that are focused on Mexico with the Mexican Health and Aging Study. And then two on Puerto Rico, which are using the Puerto Rican Elderly Study. But they're looking at things from a slightly different angle. So our first presenter is Julio Fernandez Vial. Vial. Vial. Sorry. Spanish is not that great. So if you go ahead and get your thing lined up there, we're going to try to get each of our presenters to go through in 12 to 13 minutes, and maybe as they're switching, we'll try to take a question or two, and hopefully we'll have a little time at the end as well for questions. People I suppose can start getting drinks and stuff too at the... Okay, go ahead. Hi. I am Julio Fernandez Vial from the Mexican Institute of Heriatrics. I am a human nutrition student and I did my special service there as an investigation. To start, I would like to talk about the challenge we're facing and the reason why we are here. Aging and the demographic transition. We expect from Mexico a change in which the older adult population will double by 2013 with an approximated working rate of 3.8 per year. And with this comes new challenges. Population rate... Aging raises a number of questions related to health, and what can we expect? Well, as an effort to understand these needs, large epidemiological studies of a specific population have been developed such as the Mexican Health Aging Study. And to talk, if we see the transimmortality of Mexican population, then we see that chronic non-communicable diseases occupy the leading causes, and these are David Smelidus, ischemic heart disease, cancer, and stroke. What can we expect of the current adult population in 20 years? Well, if we see that these are adults, these adults will grow. They won't magically disappear and appear, no, as an older adult. So if the transimmortality and morbidity keep going, we're going to have the same profile of David Smelidus, ischemic heart disease, cancer, and stroke, but as an older adult, which will be a problem. How can we modify these trends? Well, it's not that easy, but healthy habits can work. If we adopt healthy habits, we can assume our health will improve, and we can even prevent and manage chronic non-communicable diseases under further evolution. How do healthy habits work? It's not as simple as do exercise, eat well, don't smoke. We all make choices. For example, we either have our coffee with black or with six tablespoons of sugar, maybe even we have cookies or even a cigarette. So before action, there are steps like beliefs and motivation. Champion and Skinner designed a health belief model which was designed to predict health-related behaviors that in turn are supposed to translate into actions that will improve health. This model is complex and has several components, like barriers, perceived threats, self-efficiency, susceptibility, and benefits to a behavior. The healthy habits beliefs, moreover, depends on several factors, like social, economic, school attendance, and morbidity. In older adults, we can see this as the adaptation to a new biological and social process, which depends on the previous and new information and the personal and environmental context. Well, we had a research question. How do the beliefs in healthy habits affect the mortality of Mexican older adults? Well, this is a secondary analysis of the Mexican Health and Aging study, a national cohort study representative of about 13 million Mexicans, older adults, and consisting of four ways. 2001 baseline, 2003, and 2012 follow-ups, and 2015, which by the time we did this, it was improved. From the original sample of 15,186 older adults, the final sample after studying those younger than 60 years old and with missing data, we had 6,368 older adults. Well, the variables. The healthy habits beliefs concert was designed with the question, do you think that a person of your age can improve his or her health through regular exercise, balanced diet, or by stop smoking? It was the section beliefs question D20. Immortality was measured in months and reported by Next of King. Something important is that we censored the first deaths of the year because of the reason that those deaths could not have the effect of healthy habits. And other variables were sociodemographic, like age, sex, marital status, years of in school, smoking status, comorbidities, cognitive decline. So reported physical activity, depression, locus of control, and activities of a daily living. And the results from a total of 6,368 older adults, the mean age at baseline, was 69.07 with a standard deviation of 7.31. And as we can see, the majority of the population was married, had few years in school, and hadn't exercised in the last two years. The majority of the population doesn't smoke currently, and the number of comorbidities is one. Well, the majority of the population had a belief that exercise, balanced diet, or stop smoking can change health status. It was the 91%. And as the results, well, mortality was associated with age being male, being single, less years of schooling, and had not exercised in the past two years. Also was associated with cognitive decline, depression, difficulty to perform activities daily living, and lower score of locus of control, which means an external locus of control. All of these are previously known variables to mortality. And we have a couple of years survival course. As we can see, the effect takes time to start, and the gap starts to get bigger by the 15 months sort of. And after that, the clear color is those who don't believe that healthy habits could modify their health. And the black one is those who do believe. We can see there's a branch, and after approximately 150 months, the effect stopped working, but the gap remained. Furthermore, we did a conservation model, which was adjusted for the variables that I mentioned before, that are previously known to have an effect on mortality. The adjustment model had a p-value of 0.001, and the adjustment model had a p-value of less than 0.001. This is really important because the hazard ratio, we can establish that there was a protective effect of believing in healthy habits. And as a conclusion, according to our analysis of this national sample, believing the potential of health-related activities to improve health has a protective effect on mortality, even after adjustment for other non-variables that can affect mortality, the association precise. Well, mortality is not simple. We all make choices, and our choices reflect on social and personal context, and the way we cope with disease and health risks. We also hypothesize that this is because of time-orison. Time-orison is established as a subjective remind time of life. The social and emotional selectivity theory proposed by Carlson said, established that other adults with short time-orison probably do not believe that health-related actions can improve his health or his own life, because the social and emotional selectivity theory also proposes that other adults with short life-orisons tend to prefer more positive information than negative one, like we can consider even these as threats. So we can also say that healthy habits could diminish the impact of chronic disease and aging, modifying the morbidity and therefore mortality, and even in those diseases, there could be improvement or slow in the progress of the disease. And this can be a window of opportunity for public policy. It has been stated that a theoretical framework can improve the effectiveness of programs. Since the subjective perception of health improvement by healthy habits studied here has an effect on mortality of survival, even after modifying for other variables, we can therefore promote the belief and assure the benefits of a behavior, regardless of age and morbidity, and work on the motivation process as well as recommendations. More than an empirical process or a sort of bossing around, we should motivate the people to understand that anyone can have healthy habits, and that those healthy habits can modify, even if they won't cure diabetes, they won't cure ischemic disease. They can help and prevent the further evolution of that. That was the acknowledgement. I would like to thank you the University of Texas in San Antonio and the Red Temática del Vecicimento Salud y Desarrollo Social en Mexico for the grant to be here. Thank you. Anybody have a quick question or two for him while we're getting the next presentation now? David X. Marquez. Since your variable is on health belief change, if you believe that health beliefs are important, so why focus on mortality and not, for example, focus on change over time, since you work with longitudinal data, for example, if they stop smoking or if they stop drinking, since we know that these variables are important for mortality instead of having mortality as your outcome. That's really interesting. We didn't have the endpoint different because those endpoints aren't as common, and because if we use mortality we can have, as we see, an almost 11 years window to check how mortality is important. So I think it was a correct decision. Perhaps we can do it, and thanks for the comment. Another question over here. Thank you for a very interesting presentation. I just had one question about on your Kaplan-Meier curves. You have the scale that goes out to 200 months or so, and I was wondering how there are mortality events 200 months after the study started, because I think the latest would be 143 or so, which would be 12 months after the study began. And then I noticed that towards the end of the Kaplan-Meier curves the lines stayed straight for a while, so I don't think there was any mortality going on there. Do you know what happened there? I'm sure it wouldn't affect the results of your main variable because the lines stay separated throughout follow-up. As I said before, we only use until 2012, so perhaps the last months, I'm not sure really, were not studied. Okay, so let's give another round of applause here. Thank you. Now Kaplan Press is going to present on Health Status of Older Island Puerto Rico. Good afternoon, everyone. My name is Katharine Perez, and I am a third year doctoral student in gerontology at the University of Southern California. I actually just took my quals last week, so yay, seeing all social demography, methods, social culture variability, and stratification. One thing I find striking as a Latina and Puerto Rican researcher is that when we think about and talk about Latino health in the United States, we tend to overlook one large population of Latinos, and that is island Puerto Ricans. All the Puerto Ricans in the U.S. are considered the second largest group of Latinos. That would actually not be the case if we included island Puerto Ricans as a separate Latino category. They would actually constitute the second largest group of Latinos whom we don't know very much about in aging and health research. So, well, since 1952, with this residence having U.S. citizenship since 1917. With U.S. citizenship, Puerto Ricans can move freely from the island to the U.S. and back, which actually distinguishes itself from other Latin American countries who do not have this free-flowing migration. As of 2010, they had a population size of 3.7 million with the life expectancy of about 76 years, which is shorter than the life expectancy here in the U.S. The elderly population in Puerto Rico is rapidly aging. This graph shows the number of elderly ages 50 and over by gender, and you can see that from 1960 to 2010, the female population has increased by 284 percent, and for males, it's grown up to almost 200 percent. In this graph, I show the percentages of the elderly population ages 15 over by country and gender to give some context as to how quickly the Puerto Rican population is aging compared to the rest of the U.S. It is remarkable that it's only taking 50 years for Puerto Ricans to catch up to the elderly population that is currently right now in the U.S. Meaning that there are now more elderly in Puerto Rico now, and more now than ever before, and we know relatively little about their health statuses, including their comparisons to the U.S. And this matters because we know aging is associated with morbidity and disability and maybe a looming economic and social burden, especially not for Puerto Ricans who are facing the economic crisis of the Plexi Island. Although we do not know for certain how the health of aging Puerto Ricans should look like, but there are macro and micro-level contexts of aging that can give us some insight into the aging Puerto Rican experience. For instance, it is possible that Puerto Rican elderly may have worse health in the U.S. The older adults in Puerto Rico grew up in a different epidemiological world compared to the U.S. In Puerto Rico, there were health interventions to eradicate vector-borne diseases at the same time it was going on in the U.S. But there were most structural and infrastructural changes or improvements that actually stalled the mortality transition by a decade. For those who are familiar with the Barker hypothesis and wear and tear theories of aging, we know that these exposures to stochastic events early in the life course such as exposure to inflammation, communicable diseases, and poverty increased their risk for early onset of chronic diseases in adulthood and early mortality. Also, throughout most of his history, residents of the island have generally been less educated and poor, and Puerto Rico has always had a higher poverty rate, even higher than those that are characterized with the individual southern states in the U.S., which we know are factors that are associated with worse health. It could also be the case that Puerto Ricans have similar health profiles as the U.S. Puerto Rico operates under the same macro-structures as the U.S., meaning that, despite its commonwealth status, Puerto Ricans are subject to the Constitution, federal law, pay taxes, and receive similar entitled benefits as other Americans. And events that do happen in the U.S. also have a direct impact on island Puerto Ricans, such as their own recession, budgetary allocations, and social welfare programs to only name a few. It could also be the case that selective migration may have an effect as well, since initially her theory of Puerto Ricans might migrate to the U.S. It is possible that after living in the U.S. for some time, they adopt these bad health behaviors, and when they become sick, they have a greater incentive to migrate back home where they are stronger from their support networks. So it might have that cancelling out effect that we never get to see, which we call the salmon buyers, for the other Latino populations, but since Puerto Ricans can actually be captured in observation, it could be the case that these healthier Puerto Ricans that migrate, they might come back and they just make it look like they're the same. And then lastly, Puerto Ricans may actually have better health in the U.S. Asian Puerto Ricans do not live in the options environments that characterize the U.S. Generally, the U.S. has worse health compared to other high income countries, which have been attributed to bad health behaviors, racism, and pollution to only name a few. And Puerto Ricans also have socio-cultural and psychosocial resources such as feminism, social support, and ethnic pride that should promote better health. Puerto Rico's not called that isla del encanto for nothing, which means the island of enchantment. So based on these factors, how should Puerto Rican health look like? In order to answer this question, we used data from the 2002 Puerto Rican Health elderly health conditions project, also known as PRECO, and we used the 2002 Health and Retirement Study, or the HRS. These are two representative studies of the respective populations, PRECO, precocious cross-sectional two-way panel study of the non-institutionalized island Puerto Rican population, age 60 and over, with over samples of people of African descent and individuals over 80. All the interviews there are conducted in Spanish. The HRS is a longitudinal health interview survey of a cohort of adults age 51 and older in the U.S. that is conducted by annually with over samples of minorities and residents of Florida. Interviews there are conducted either in English or Spanish. For this sample, we restricted our sample to include older adults ages 60 and older and include respondents with non-missing health data. So this includes 10,679 HRS whites, 1,160 HRS Latinos, and 4,389 Island Puerto Rican from PRECO. We decided to choose these two U.S. groups because we are using whites as a benchmark to help characterize the Puerto Rican population, and we also include Latinos to see if since Puerto Ricans are considered Latino groups, let's see if they actually look like other Latinos in the U.S. Because we are interested in obtaining complete health profiles of Puerto Ricans, we look at health as a multi-dimensional concept, and we use three domains to describe health in this study, chronic conditions, disability, and self-rated health. We use chronic disease and disability because these conditions are primarily present in elderly populations and we use self-rated health as an assessment of an individual's health, which has been consistently found as a validated measure for assessing a person's health. We use hypertension, heart disease, stroke, cancer, lung disease, and diabetes to characterize chronic conditions. We assess disability as any limitations in activities of daily living, which include bathing, eating, dressing, walking across the room and getting in and out of bed, and self-rated health is assessed by whether the responder reports for it to poor self-rated health, and we also know that since each of these health conditions operate differently by gender, this study also investigates whether these dimensions of health are also affected by gender. For these eight health outcomes, we use weighted logistic regression models controlling for age, gender, education, medical status, insurance status, and health behaviors, and using the estimates from these logistic regression models, we calculate our predictive probabilities with 95% gender group for the purposes of this presentation in order to illustrate the health differentials. So, just to kind of situate and kind of orient you as to how the results are going to be shown, so first I'm going to go through the patterns of males by the first three diseases and then I'll show females. So for these graphs, whites are going to be characterized by the blue color. So, on this graph, you'll see that there is no variability in hypertension among men. Here, you'll see that Puerto Rican men are less likely to have heart disease compared to whites and have a larger gap than there's a larger gap between Puerto Rican males and white males than it is between Latino males and white males. With diabetes, there isn't much difference between Puerto Rican men and Puerto Rican women are much larger than the differences that you see between whites in general. Puerto Rican women have less heart disease than whites in general. So, here you can actually see that Puerto Rican women have more hypertension compared to any other group. Also, I want to note the difference in hypertension between Puerto Rican men and Puerto Rican women have more heart disease than white females and look actually more like they're Latino counterparts. Here, the gender gap is a lot closer together than the gender gap that is observed of long whites. Puerto Rican females have more diabetes than white females and look actually more like that thing is here too. Here, you can see that the gender gap is more closed among Puerto Rican than whites. Here, for stroke, Puerto Rican males here are also less likely to have cancer compared to whites. Here, Puerto Rican females are less likely to have stroke than whites. Although you can see that Puerto Rican females have similar patterns of lung disease here as white males and white females, there is actually a really large gender gap between Puerto Rican women and Puerto Rican men compared to what you see among whites here. And here, Puerto Rican women also are less likely to have cancer compared to whites. When it comes to any ADL limitations, Puerto Rican men are less likely to report any compared to whites and white females. Puerto Rican males are also less likely to report compared to poor, so-for-rated health. And here, although women, Puerto Rican women are less likely to report any ADL limitation, the gender gap between Puerto Rican men and Puerto Rican women are a little bit larger than the gaps that you see among men. And here, Puerto Rican women and Puerto Rican women are having a more difficult time than their male counterparts. And then here, Puerto Rican women have the same likelihood of reporting compared to poor, so-for-rated health as their white counterparts, but again, there is this gender disparity among Puerto Rican. So I know I kind of went through a lot of results really quickly, so I'm just going to try to recap all of this in one slide, or two. So, overall, we find that Puerto Rican males have actually healthier profiles compared to the white counterpart. Puerto Rican men are less likely to have heart disease, stroke, lung disease, cancer, less likely to report any ADL limitation and less likely to have negative ratings of their so-for-rated health. Whereas, Puerto Rican women are less likely to have heart disease, stroke, and cancer. Among Puerto Rican males, so, women are doing worse. When it comes to hypertension, heart disease, lung disease, any ADL, or stroke-related health, women are doing much, much worse than their male counterparts. So, why are Puerto Rican males doing so well? Well, I have some hypotheses that haven't directed test disease, but these are just thoughts I'm having. So one of the reasons why there might be this Puerto Rican male health advantage is that there might be a selectivity for the out-migration of male workers that do not return to the island are driven by those who are more of a lower socioeconomic background who tend to leave the island to go look for work in the mainland. And also, I have a paper that I actually have under review right now that shows using NHIN data that Puerto Rican migrants from the island, it doesn't matter what your time of duration is in the U.S., you are going to be more sicker, you're more likely to be sick and also, there are theories of cumulative advantage. Males are more likely to have access to education, better employment opportunities, more likely to have continuous employment and less accumulation which we all know factors are associated with better health. And the Puerto Rican female health disadvantage can be characterized by this theory that we know as a double jeopardy or multiple hazard hypothesis. So Puerto Rican females on the island are more likely to be headed, have more female headed health and they're burdened with caretaking. They're actually just experienced more discontinuity in their life course in labor which gives them more economic disadvantage and they're also historical insults. For those who are not familiar with Puerto Rican history, there was a period where Puerto Rican women went through undergone forced sterilization as a result of some of the eugenics thinking that was implemented by the U.S. government. So I'm sure that has some effect on the long-term assessments of health. And as I move forward with, you know, digging more digging into Preca, I'm hoping to be able to find some of these other sources of variability that explain this gender disparity and my time is up. So, thank you. Okay. Got any questions while we're getting set up? Brian, if you want to go ahead and get your presentation. I have a lot of results. I find them interesting but I'm wondering do you, what hole do you have on the health care delivery systems in contrast one to the other? U.S. and Puerto Rico? Yeah. So those are parts of data analysis I'm sort of trying to get through. What I find very interesting is that at least in the island of Puerto Rico, there is like almost universal health care for these older adults. But I don't know yet, you know, their patterns of health utilization. So even though you might have it could be the case that maybe not enough people are going to see the doctor, maybe you have this like selectivity effect too, maybe you know that's why we're seeing probably healthier Puerto Ricans right now might not be the case. But I would have to look into the variables of utilization to be able to answer that. At the beginning of your presentation, you sort of were making the case that Puerto Ricans look probably different than other Latinos and that was sort of a group from Latinos. I saw differences between Puerto Ricans and US whites. But across the bars, I did not see large differences in the difference between Latinos and Puerto Ricans. In some cases, the levels were likely different. But once we considered a confidence interval, they pretty much overlap. Yeah, they all pretty much overlap. So now my question is, are they really profiles alone? Yeah, they look the same, right? But the problem is we don't know what's driving the health behind the Puerto Rican population versus the Latino population. Because we don't like the discussions of the Hispanic Paradox, having that selective migrating and the salmon buyers. And these are things that we have to take into account. Because yes, even though they look the same, but these groups all have different values, it doesn't mean that the same causal factors are contributing to these similar patterns. If that makes sense. Okay, thank you. Catherine, that's great. I'm sorry, but we're going to move on. Because we're kind of getting tight on time here. So, Brian Downer is going to keep with the Puerto Rican theme here focusing on diabetes, depression, disability, and mortality. Thank you. Good evening, everyone. It's a great opportunity. And also, I'd like to thank my co-authors, Dr. Michael Crowe from University of Alabama, Birmingham, who's been doing a lot of work with Preco, and then also Dr. Coak, who's many of us know. So my background slides got a little bit shorter, and some methods slides got a little bit shorter since we had a nice summary of the Preco data in Puerto Rico in general. We'll talk a little bit about aging in Puerto Rico before we go on to the methods and the results. And I have a few concluding remarks. Briefly, as we all know and have heard repeatedly, type 2 diabetes is a significant public health concern, especially for Hispanic populations and other minority populations. Also, diabetes is often comorbid with depression and individuals with diabetes are more likely to develop depression or depressive symptoms. Also, there's been some research that suggests that individuals who have comorbid depression and diabetes have severe diabetes and are at greater risk for diabetes-related complications. In the slide here, this is some data from the Hispanic Epis kind of supporting that notion. There's also been a fairly extensive body literature looking at the relationship between depression, diabetes and other health-related outcomes, including ADL disability and mortality. And again, the same kind of conclusion seems to apply. This is from the same Hispanic group showing that individuals with both diabetes and depressive symptoms are at greater risk for ADL disability and mortality compared to individuals with diabetes alone or depression alone. I want to talk about some of the background about Puerto Rico. I would like to bring everyone's attention to the ongoing popular crisis that's currently happening in Puerto Rico. In 2013, there were approximately 3.7 million total population. There were just under 10,000 physicians and particularly relevant to our conversation only being 45 geriatric medicine physicians. In addition, in 2014, roughly 360 physicians left the island of Puerto Rico to come to the mainland of the United States. And these potential challenges are compounded by the fact that Puerto Rico receives less federal funding than other 50 states in regards to health care. The population distribution in terms of age, as was mentioned previously, Puerto Rico is aging quite rapidly, being brought on by declines in birth rate, but being accelerated by young adults who are working age leaving for the United States seeking employment opportunities. In terms of the percentage of the population that's older than 65 and older than 85, Puerto Rico is slightly older than the mainland of the United States, but it's fairly comparable. So for the site I'm presenting with you today, the first objective was to simply examine the relationship between diabetes, depression, and the likelihood for developing ideal disability and mortality. And then also looking to see if there's differences in these relationships after accounting for disease severity determine based on if an individual is receiving treatment for their diabetes and their depression. So, and that's a previous presentation, we used data from the baseline survey which was in 2002, 2003, and then also used the follow-up survey which was completed in 2006 and 2007. For this analysis, we used the 4,200 individuals from the baseline excluding those who received a proxy interview, missing information for covariates. The mortality sample was just almost 3,500 and then we subsequently excluded individuals who were deceased to also follow-up and also who were ADL disabled at baseline for the ADL analysis. For our covariates we used demographic characteristics, self-reported health conditions, and we also used cognitive functioning. For diabetes, we used self-reported diabetes which participants responded to ever having been diagnosed by a physician. Those individuals who also reported that they were receiving insulin shots to control their diabetes were this information was used to classify participants as non-diabetic, diabetic, but not reporting to use insulin and then diabetic and reporting to use insulin. We used a similar strategy for categorizing the depressive symptoms. So individuals who had evidence for high depressive symptoms but were not did not report seeking psychiatric or psychological treatment were categorized as having high depressive symptoms and then we had our third category high depressive symptoms and also seeking psychological or psychiatric treatment. Just real quickly, for the activities that they live in we used a similar scheme to the previous presentation where anyone was disabled and one or more ADLs they were qualified as being ADL disabled. I will point out that the way that the question is worried for the preco is the difficulty has to be due to a health problem and then participants are also asked to not consider temporary difficulties that they may be having if they have recently had a surgery or some other acute health event. For this statistical analysis fairly straightforward looking at characteristics based on mortality status and then using multivariable logistic regression models. So here so the mean age of our these are all the baseline characteristics just over 71 years of age and on average completing eight years of education majority female and about 41% of our final sample was Mary. Now for some of the health characteristics so they see that there is a fairly high prevalence of self reported diabetes population and then also over half the population sample population reporting self reported hypertension and then lower prevalence for heart attack and stroke. Also about half of the sample reported having been diagnosed with arthritis and then 20% had high evidence for high depressive symptoms and 14% were qualified as being ADL disabled. So here's the results for simply looking at if a person was diagnosed of having diabetes and no diabetes and then depressive symptoms or no depressive symptoms and as you can see here my orientation got off a little bit but both diabetes and high depressive symptoms were associated with higher odds for ADL disability and mortality. And then this model this is taking into account the different treatment and it's almost receiving treatment for diabetes or not and here we can see the difference based on treatment status. So for ADL disability individuals who had self reported diabetes but did not receive insulin they did not have significantly higher odds for ADL disability but they did have higher odds for mortality and then even higher odds for those who had self reported diabetes and were also taking insulin. We see a somewhat similar pattern for depression with the highest odds for ADL disability and mortality being for those who had high depressive symptoms and were also seeking treatment for those symptoms. So there's a few points to conclude so I talked about this a little bit in the introduction but I think it bears repeating again so there hasn't been a lot of academic literature published on healthcare in Puerto Rico but I was able to find a nice summary article written by a physician actually at the University of Louisville that I encourage everyone to read because a nice overview of the healthcare system in Puerto Rico and I think it's important to keep in mind when interpreting the results from the study within the context of a healthcare crisis so being able to treat and manage diabetes and depression is going to become even more challenging when you put it into the context that there's a lack of physicians and there's a lack of resources to treat these conditions which kind of led me to think a little bit about is using treatment status as a measure of disease severity is that maybe the best approach for this population so a lot of studies have used diabetes treatment as an indicator for disease severity but it may not necessarily be the case for the best approach for Puerto Rico because not receiving treatment may not reflect less disease severity it may be a consequence of limited access not being able to afford treatment and so I think it's important to keep that in mind when interpreting some of these results some of the limitations so again the self-employed diabetes requiring that you be diagnosed by physicians so undiagnosed cases of diabetes won't be correctly classified similarly not the clinical diagnosis of depression Preco does include a variable in which participants are asked if a physician is told that they're prone to depression but it's not entirely clear with that information that's telling us and then also some of the individuals in the follow-up interview depending on what the cognitive functioning is they may not receive full interview and so given the relationship between cognitive impairment and ADL disability we may have a slightly different follow-up sample so again just to summarize the findings we've replicated what's been observed in many other reports with diabetes and depression being associated with an increased risk for ADL disability mortality with some evidence suggests that treatment may be related to that risk and also again I think these results highlight the importance of older adults with diabetes and or depression especially important to go to be able to find way to access appropriate treatment and management of these conditions. Thank you. Any questions for Brian? Brian, good job. Very interesting. I wish I could make a presentation like you I need to study talking about that. So my question is about just since we went through those models so quickly because there's a lot of information the diabetes model and the depression model were those separate models or were they were they co-buried together or did you look at the two like for those who had both and just to see if there is like this additive effect or you know what I'm getting at on that. Actually I'm glad you mentioned that Ben so the diabetes and depression were both included in the same model. Like I said that's the concise answer. I had a question. Do you do you have any cause-specific mortality data or are you able to kind of determine? I don't believe so. So that would be an answer for a question for Dr. Crow. My understanding is there's not cause-specific mortality. So this mortality data hasn't been linked with the National Death Index or anything. So I think they're in the best position to be able to make that linkage which I believe would give them information on cause of death. Because I just think looking at I mean depression you might expect there to be perhaps higher rates of suicide and maybe for diabetes some other kind of associated cause of death as well. So that would be an interesting experiment. Yeah. Any other questions? Okay. Well our next presenter is Jacqueline Contrera. Take it away. Hello. I'm the last presenter so I know that you are tired and I also am but I'll be concise and hopefully we'll learn something together. I'm a PhD student I just started my second year at the Department of Preventive Medicine and Community Health at UTMB and I developed this work it's actually an ongoing project with my two mentors Dr. Sopna Call and Dr. Rebecca Long and our project is about healthcare expenditures and utilization among Mexican older adults. But before I explained our project and our rationale I would like to contextualize you on the healthcare system of Mexico. I know we talked today about that and also you are probably better experienced than I am but let's review that together. So let's take the year of 2001 for example or the year of the 2000s. What happened in this period in Mexico is that people that were formerly employed either with public or private employees were insured small proportion of the population was insured through private insurance but nearly half of the population was uninsured and 53% of total health spending was without pocket costs and these factors motivated a healthcare reform that specifically happened in the year 2003 and it came specifically in a form of a new public health insurance called Seguro Popular which means popular insurance in English and its main goal was to provide insurance to those that were previously uninsured so basically people that worked in the informal sector of the workplace or people that lived in the rural areas basically provided insurance for these individuals and also its second goal, major goal was to reduce out of pocket expenditures so what have we seen so far after the healthcare reform? Let's take the 2010 year as an example here by then there was a 20 to 30% reduction in total proportion of uninsured and a 54% reduction in total out of pocket expenditures and along with the change in the healthcare system we all heard that there was an accelerating aging in Mexico and in Latin America and instead of restating that I would like just to show you one example so in 2010 6% of the Mexican population was above 65 years old and the estimations that by 2036 this number will reach 15% so that's a 26 year span if you compare to other developed countries for example the United States took 69 years to go from 6% to 15% of older adults and it took France 115 years so that's just one example of so many that you also presented a show of the rapid aging in Mexico and this rapid aging in the population still faces other challenges where you have a growing problem with chronic diseases where still infectious diseases are still very prominent and as a consequence the elder individuals have the highest utilization of ambulatory care and hospitalizations and also have the highest proportion of total out-of-pocket expenditures so that's our rationale right there this population is growing these people are using more services these people are paying more out-of-pocket so we need to study them and we need to study who they are so that is our objective our first objective is to cross-sectionally compare just for us to describe the health care utilization and expenditures in 2001 in a moment before the health care reform in 2012 in a moment after the 2003 reform and then the second objective to identify the video characteristics for example comorbidities or insurance status that are associated with health care use and expenditures in Mexican older adults so to achieve our objectives we use the NS you heard a lot about it today so let's just look at the figure you can see the four waves of NS starting in 2001 then in 2003 and then in 2012 where a refreshed sample was added with younger individuals and the last one in 2015 for our study we utilize the waves one and three as the moment before and after Seguro Popular to compare a sample size of 12,000 direct interviews with a population of 13,000 direct interviews in the future we would like the same individuals longitudinally but first we want to describe this population before and after and see if there are any differences first objective our main outcome variables were if these older adults reported any health care services in the previous year so in the end we have information about if they spent any nights in the hospital or if they had any medical or outpatient visits and our second outcome is that what were the 2001 Mexican pesos 2012 Mexican pesos to adjust for inflation so these numbers are comparable now and to answer our second objective and see the individual characteristics associated with higher use or higher expenditure we looked at different variables sex age merit status education insurance status the first two insurances in East IUC are public insurances in Mexico that ensure employees and public employees respectively in Seguro Popular we only have information on 2012 but that's the new insurance that we're looking at if they had private or other insurance if they were uninsured finally we looked at a comorbidity score if these older adults reported 0, 1, 2, 3 or more chronic diseases okay so now moving on to our analysis to answer our first objective just compare we just compared the differences between 2001 and 2012 of the ANHAS data using a bivariate analysis but then to identify the covariates associated with expenditures and use we also utilized a two-part regression model as the other presenter here today also utilized this is a very good model to model this kind of data of healthcare use and healthcare expenditure so in the first part we use a logistic regression model so we just compare people that use services versus people that did not use services and people that paid for some amount out of pocket versus people that did not pay for some amount out of pocket and then our second part model was a GLM model based on the probability of having at least some service use or based on the probability of paying at least some amount out of pocket okay so I showed you our rationale our objectives our analysis and I selected just some key results to be concise and show descriptive results from 2001 and 2012 in 2001 the mean age was 62 years old and in 2012 65 years old you only see a three year difference but remember that in 2012 some younger individuals were added to the sample that's why the difference is not that large interestingly interestingly too the people that recorded more than seven years of education increased from 2001 to 2012 but that is also could be driven to the younger individuals now in the 2012 sample more individuals reported zero chronic diseases again this self-reported data but still shows some positive results but perhaps my favorite result is this one that the proportion of the uninsured dramatically decreased from 2001 to 2012 you can see the uninsured in the blue line so in 2012 there were 45% of older adults uninsured in Mexico in 2001 excuse me and then by 2012 the number decreased to 15% so that is a great result that I observed in our data and you can see Seguro Popular there in the purple X around 31% of these older adults are now insured through the new health insurance Seguro Popular let's see how our main outcomes change between those two years if we look at the proportion of the elderly with any health characterization in the previous year we see that proportionally older adults proportionally spend more time in the hospital and have more medical visits in 2001 which is in blue then in 2012 in red they are using more but if you look at the mean number of nights in the hospital they are actually staying less nights in the hospital and having more medical visits so that just shows that they are now using more services but what our next slide shows is what is interesting about this change is that if you look at any out-of-pocket expenditure in the previous year proportion the proportion of payers decreased from 2001 to 2012 so we use increased but the proportion of people that are actually paying out-of-pocket for these services decreased even though the mean number the mean amount that they are paying that's why you've seen the boxes from 2001 to 2012 increased that is just due to normal health care increase costs that we see everywhere but proportionally fewer people are paying out-of-pocket and that is a great result that we see from a now more insured population result I selected just one my varied analysis for you to see the average of total out-of-pocket expenditures in the previous year by comorbidity score what do I want to show with this slide that there is a positive correlation between average expenditures and comorbidity score take the year of 2012 for example in red if you look at people with zero comorbidities then in average 4,000 Mexican pesos out-of-pocket but if you look at people with three or more diseases they spend 10,000 more than double that is a high disparity that we further analyze in our regression models that's what I'm going to show you now so in this multi-variable analysis you're not looking at coefficients but you're looking at the marginal effect so if you see a plus sign it means that in average these people used more or spent more than their counterparts in the reference group for that variable if you see a minus it means that they spent less or used less than their counterparts for example what do we see for older adults the oldest old those above seven years old we see that they use more and spend more compared to the younger between 50 to 59 years old and I would like to highlight that you looked at insurance status through IMS IMS is a public health insurance in Mexico that ensures private employees so you see that they're more likely to use services they're more likely to spend nights in the hospital and have medical visits but they spend less out of pocket and that's because they are now insured if you look at Seguro Popular in 2012 you see that people spend more nights in the hospital they have more medical visits and even though the out of pocket cost is not significant it is in the expected direction and maybe in the future we will observe the benefit but what I really want to highlight here is the comorbidisk across all categories people are more likely to spend and more likely to use services and I actually want to show you some numbers so take the year of 2012 on your right for example people that have one disease spend an extra in average 600 extra Mexican pesos than people that are healthier and report zero diseases and people that have three or more diseases spend in average 2,000 Mexican pesos extra than people that report zero diseases and this shows the high disparity between healthier and the seeker and the financial burden that the seeker individuals face as an older adult facing all these financial charges out of pocket so I showed you some selective results but what are main conclusions from this study comparing before and after the healthcare reform in Mexico show us are very simple so I put very simple conclusions insurance coverage increase that's what we need to take home utilizations of services increase but the proportion of people with any out of pocket expenditure decrease that's what we expect from this population that are now in insurance however those with more comorbidities the oldest old and those without insurance still incur the greatest out of pocket expenditures and this is our next step my first next step is to look at longitudinally and compare the same people in 2001 2012 how that changed and also understand their health behaviors what are their specific diseases do they have disabilities how their social demographic distribution to then be able to assess this very vulnerable population so before I finish I would like to thank my mentors Dr. Rebecca Long and Dr. Sapna Kall and acknowledge the funding received by NIS by NIN okay any questions for Jacqueline thank you Jacqueline thank you very good I was wondering to what extent you think I was wondering if you analyze any rural urban differences yet because I'm thinking maybe in the rural areas where this is more prevalent the use of the Segura Popular so yes we look the location size so comparing size how serious less than 100,000 people and more than 100,000 people but at the end these results were not significant in the multivariable analysis so they didn't significantly change from between people that lived in the more rural or more urban area but that's definitely something that I want to at least look at it one more time because that was also one of our hypothesis thank you any questions any questions I don't know I don't know if you had a chance I still remembered your question and I cut you off is there a large disparity between the education level of men and women in Puerto Rico? it's really funny that you ask that question because in the descriptive statistics men and women look the same across the different educational gradients so there weren't like any significant biorariate differences by gender and the other part of the question was the men that are leaving the island and then coming back what is socioeconomic level are they across the board or the lower economic level? so my hypothesis is that there tends to be more of lower socioeconomic status and the problem that I cannot verify that is because if they don't come back to the island I'm not going to be able to get that information because I didn't come back but they are such a small group of people that I can't really look at any variability within that but my assumption is that the more lower socioeconomic status people leave the island because if the island has already almost a 50% poverty rate and there's no jobs there's unemployment you need to look for a job somewhere to survive so you gotta go from what I've read the like around a high school education and lower tend to be the individuals that are leaving the island to the United States so it fits with what you're thinking are women leaving as well or is it just the men? I'm not I don't know that for certain but my hypothesis is men just because given the Puerto Rican history of migration movements related to work so when operation bootstrap states that was mainly comprised of pretty much all men and then there's been a lot of worker migrations from the military taking Puerto Ricans to the mainland which were also mainly consistent of men okay so one more question we need to wrap it up just for Jacqueline and I'm sorry if you presented it and I missed it so probably because probably I did but so is there a way to separate what's happening between high insurance which in Mexico doesn't really mean much especially for public I mean you can be insured but what real access do you have to services so is there a way to separate that against help just help I mean are they spending more nights in hospital because even if they are younger even with the sampling with a young I mean they could be 50 but still diabetes hypertensive and so is my question is could you really separate the effect of being just with a lot of more comorbidities than so what are we seeing is it because they are more you know now insured because they are sicker yes I think that's how that was our question too and then moving forward now that's what you want to understand so who are these individuals that are spending more because this was basically just a descriptive first we just wanted to see if there was going to be a change and now that we actually saw this and we see that these individuals now have more comorbidities we want to understand who they are are they insured you know see if these individuals are insured see what are the age of these individuals if they have disabilities or not so I think this is definitely going to be my next step before I present this paper in written form sorry I don't wanted to give each of our emerging scholars a certificate from UTSA so I can do this with my we'll get a picture here I think I was you know I've certainly read through their abstracts but really I'm very impressed with the work that they're doing and certainly their ability to organize and present so please let's give them another round of applause and then I think there's time for a stand would you like them to do a come down or thank you now I think it's time for the award ceremony and is that going to be in here and open that's not an open bar is it and it's in order so let's adjourn into this space thank you