Exposure to a Natural Disaster and Long-term Cognition

Jump to Transcript

In this episode, Matt and Lauren speak with Drs. Elizabeth Frankenberg at the University of North Carolina at Chapel Hill and Duncan Thomas at Duke University about their research which focuses on understanding how survival and physical health evolve after exposure to large scale shocks like the 2004 Indian Ocean Earthquake and Tsunami. They will specifically speak about the STAR Study and what they’re learning about long-term effects on cognitive outcomes.   

More resources 

Study of the Tsunami Aftermath and Recovery (STAR) 

STAR Study Design and Results 

Transcript

Matt Davis:

In a previous episode, we discussed the effects of exposure to hurricanes among older adults living with cognitive impairment who rely on consistent caregiving and healthcare. While such studies historically have focused on mortality and sometimes, morbidity, recent work has started to uncover the more insidious effects of natural disasters. A large-scale disaster can severely damage lives and livelihoods exposing entire communities to stress for many years, particularly in low and middle-income countries that have fewer resources, the effects of stress can be wide-ranging, extending even into cognitive outcomes.

Today, we're going to speak with a research team that's studying the long-term impacts of the 2004 Indian Ocean earthquake and tsunami on memory function. Welcome to Minding Memory, a podcast devoted to exploring research on Alzheimer's disease and other related dementias. Here, we'll discuss some of the most compelling research and talk with leaders in the field about how their work is improving the detection and treatment of dementia. I'm Matt Davis.

Lauren Gerlach:

And I'm Lauren Gerlach.

Matt Davis:

We're both researchers at the University of Michigan. I have a PhD in data science.

Lauren Gerlach:

And I'm a geriatric psychiatrist who specializes in diagnosis and management of dementia.

Matt Davis:

I'll work to minimize the use of medical jargon in our discussions.

Lauren Gerlach:

And I'll make sure that the research we talk about has practical, real-world applications to people living with dementia and their care partners.

Matt Davis:

Thanks for joining us and let's get started. We're joined today by Doctors Elizabeth Frankenberg and Duncan Thomas. Dr. Frankenberg is the Cary C. Boshamer Distinguished Professor of Sociology at the University of North Carolina Chapel Hill. Dr. Thomas is a Norb F. Schaefer Distinguished Professor in the Department of Economics at Duke University. Doctors Frankenberg and Thomas have been collaborators on several large studies, including the Study of the Tsunami Aftermath and Recovery, also known as STAR for short, the Indonesia Family Life Survey and the Work and Iron Status Evaluation. Their team's work focuses in understanding how survival and physical and psychosocial health evolve after exposure to large-scale events. They're here today to speak with us about the STAR study and what they're learning about the long-term effects on cognitive outcomes. Elizabeth, Duncan, welcome to the podcast.

Elizabeth Frankenberg:

Thanks so much for having us.

Duncan Thomas:

Yes. Thank you very, very much.

Matt Davis:

So, before we jump into discussing your study, just as a quick disclaimer, in addition to being long-time scientific collaborators, you're married, right?

Elizabeth Frankenberg:

That is correct.

Matt Davis:

I think we could do an entire podcast on this topic. I've known a few married couples in the past and I have a million questions for you, but I will hold back on those questions.

Duncan Thomas:

Can't wait for that podcast.

Matt Davis:

Okay. So getting down to business here, I was wondering if you could tell us or talk a little bit about your thoughts about the causal mechanism between exposure to a disaster and long-term cognitive outcomes.

Duncan Thomas:

So I think in general, identifying causal determinants of cognitive outcomes, especially over the long term, is really a major challenge in the literature. And one of the reasons that we're doing the work we're doing now is that there is a literature that has related exposure to stress to some domains of cognitive function. So for example, we know that exposure to disaster is going to result in reduced economic resources because of all the damage from a hurricane or a tsunami or something like that, reduced social resources, either because you move and you are displaced or because family members have moved or been displaced or even people dying. And of course, linked to all of that is elevated stress. And what we know is that stress has a direct effect on cognition. It's been linked to worse memory, less attention, worse response inhibition and worse cognitive flexibility.

Most of these results are from the animal world, from animal models. There's not a lot of good causal evidence from human populations. And what's good about this literature is that for the effect of stress, from the animal models, we know that they have an adverse effect on medial prefrontal cortex and that we can conclude from that literature that there should be a causal impact on memory function possibly in human populations unless there's a lot of adaption. So we also know that stress has a causal impact on cardiometabolic dysregulation and we know that that in turn, affects cognitive performance or at least probability of dementia. So there's two potential domains from two potential pathways for there to be a causal impact on memory function. Other causal impacts on other domains of cognition, I think that is more difficult to pin down because the link to what part of the brain is being implicated is not as clear.

So what we also know is that with the tsunami, so first, we know the tsunami caused a lot of stress. The stress would be widespread from many, many different dimensions, but the tsunami also caused a huge amount of dislocation of resources, damage to resources, destruction to resources and dislocation to social resources, connections to family and friends. And all of those are likely to be rolled up into impacts that would affect cognitive function and specifically, memory. And the other reason why other domains of cognitive function may be difficult to pin down causal impacts is that of course, if after a tsunami you have to rebuild your life and re-navigate or navigate a new and novel landscape, then it's quite possible that you had to build on cognitive reserve in ways that you would not have had to do if you didn't go through the tsunami.

So in fact, you may end up having improved cognitive function in some domains like executive function because you had to pull all these things together when you were in your mid-30s or mid-40s or mid-50s when you were really planning to just carry on going as you had done for the last 10 years. So the reality is that while we have a theory, we don't know that the impacts will necessarily be deleterious, and it's really an empirical question, what are the long-term impacts of exposure to tsunami on cognition? And we don't have a ton of evidence because not many people have followed a population post large-scale disaster like a tsunami.

Matt Davis:

We're getting ahead of ourselves a little bit, but it does make me curious about the effects at different ages. It's got to be very different in terms of exposure to stress at different points in your life. You must be doing a lot with stratified analyses by age, I imagine.

Duncan Thomas:

You're absolutely right. So the stuff that I was planning to talk about here would be mostly about older adults. So when we look at memory function, honestly we do not... So we don't even measure memory function for young adults at the moment. We will in the future, but at the moment, the study we've done so far, which is relatively modest, is really people aged 45 and older, and we don't see any deficits until people reach around about 50. So they would've been exposed around about age 35, 15 years before. And so it's not that it only affects really old people, but it's only emerging among this older cohort. And it's very interesting to see if people are displaying premature cognitive aging because of the stress. But you're absolutely right, the effects get bigger and bigger as they get older and older, but they're quite modest for the younger people. So that's an open question for us that we hope to do in future work. So that'll come up at the end. But yes, we're stratifying stuff by age.

Lauren Gerlach:

I was hoping you could tell us a little bit about the study of the Tsunami Aftermath and Recovery or the STAR study. In particular, what measures are you collecting and how long has the study been going on for?

Elizabeth Frankenberg:

Sure. So we've been collecting the data for almost 20 years now, and what happened was that there was the earthquake and that triggered this enormous tsunami. And Duncan and I had worked in Indonesia with our close collaborators at an NGO called Survey Meter in Indonesia, and we'd been doing these long-term longitudinal surveys. And so we felt like this was an event of such magnitude that it was important to try to collect rigorous data over time, about how people were affected. So we were able to work with Statistics Indonesia, who was very enthusiastic about us taking one of their cross-sectional surveys done before the tsunami and converting it to a panel.

So we worked with them to make this conversion to longitudinal data collection, started in the field about five months after the tsunami. And the measures we collect have changed over time to some degree. We went into the field annually for the first five years and collected data about demographic and health outcomes, consequences of the tsunami, markers of physical and psychosocial health. Then after the first five years, we started doing the surveys every five years and did more long-term retrospectives about migration and marriage and family formation. And so we've been following these people now for 20 years, looking at how health and economic resources and psychosocial markers have evolved over time.

Lauren Gerlach:

For our listeners who may not be familiar, can you tell us a little bit about the depth of the devastation with the 2004 Indian Ocean earthquake?

Elizabeth Frankenberg:

Sure. So the earthquake itself was incredibly powerful. It was 9.3 on the Richter scale and actually shifted the magnetic location of the North Pole by a couple of centimeters. And it generated an earthquake that traveled a 1200 kilometer pathway up this rift zone. And that's what displaced all this water that slammed into coastlines around all the countries with Indian Ocean border. And the waves reached Indonesia's shores about 15 minutes after the quake. So very little time.

They were 30 meters high in some places, so they were enormous walls of water. And some of the landscapes afterwards really look like lunar landscapes. Practically everything has been removed and what was there has just been scraped away and replaced by sand. So it was immensely destructive, but the degree of destruction varies a lot from place to place and is connected to things like the topography of the ocean floor because that generates a huge variation in the height of the waves as they come ashore. And so that generated this spatial idiosyncrasies in which communities were really devastated and which had less damage.

Lauren Gerlach:

Wow. And in your work, have you considered any other historic disasters or have there been other disasters that you've looked at as well?

Elizabeth Frankenberg:

So, for Indonesia, we had looked at the economic downturn of 1998, and then we had also used a similar approach to look at the effect of the Bali terrorist bombings of 2002. And so the idea of trying to build longitudinal data from before and after a major event of some kind was something we had worked on before, but we had never worked on disasters before.

Lauren Gerlach:

And I guess I'll ask this for maybe some of our listeners who are NIH-supported or inspiring investigators who are hoping to get NIH support. Is there anything unique about NIH proposals for international studies that might be helpful for our listeners to know or maybe some of the words of wisdom for our listeners here?

Elizabeth Frankenberg:

Well, I think making the case for why a study internationally has relevance for U.S. populations as well is quite important. And so this disaster was very large-scale. The U.S. is affected by major flooding every year. And so I think those are important points to establish. We've been incredibly fortunate to have NIH support and we're very grateful for it.

Duncan Thomas:

Let me just add two things to that. So first is that one reason why I have done the vast majority of my work in low-income settings is that the variation you see in those places is so great relative to the variation you would see in the United States that there is scope for pinning down some impacts that would be very muted in a U.S. setting. And the second is that, for example, Katrina was a disaster that occurred the year after the tsunami, and there are no large-scale follow-ups of the population that was affected by Katrina.

Now, that's in part because the nature of data collection in the United States is very different, but it's also in part the difference between Statistics Indonesia reaching out to us and saying, "We have this pre-tsunami baseline of about 25,000 people who were living along the coast of Aceh, the area that was most damaged by the tsunami. We don't have the time to put together a panel survey. Would you be willing to collaborate with us to actually develop a panel survey?" That is not something that one     could do in the United States under the current structure. And so that's something that makes the possibility of having causal mechanisms at play, feasible. And that is the argument I think one wants to make to NIH, that you will contribute to science. And that's fundamentally why we should be doing this and why this is truly a scientific study. It's not easy to work in an international context, but it is incredibly rewarding for these reasons. From a scientist's point of view.

Matt Davis:

I'm so glad this came up because we see a lot of people developing grant proposals and stuff, and there's always a little bit of tension on trying to improve the health in a different country versus the broader scientific utility of what they're trying to do. And I just know it's a line that people walk.

Duncan Thomas:

I think it's important... That's a really valid and important point. I think it's important to recognize that we're not claiming that this has no relevance for Indonesia. Of course it does. And we certainly have put a lot of work into making sure that what we've learned has been transmitted to the Indonesian policymakers, decision makers and so on. But if all you're doing is improving the health of some country somewhere, the population in some place, that probably is not the kind of thing NIH is going to fund. And so you do need to make the case that it's going to be impactful on science or has a potential to be impactful on science. And I do think that if you do these kinds of studies, you have to be fully committed to contributing to the communities and to the populations that you're working with. And we are. We work with the most amazing collaborators. That's truly one of the pleasures of working with a team we work with.

Matt Davis:

That is impressive. I was going to say, it's even impressive too, to have a grant going for that long as well and keeping it renewed and everything.

Duncan Thomas:

Well, we've been working with these guys... The guy who leads this project name is Cepep Sumantri. Dr. Cepep Sumantri. We've been working with him for 30 years. This project has only been going for 20 years, or there about, 19, 18 years, whatever.

Matt Davis:

So when it comes to disasters, hurricanes, earthquakes, etc, that occur in low to middle income countries, what should our listeners know? How is it different than here in the United States?

Elizabeth Frankenberg:

So, what drives the disaster creates some of those differences. So earthquakes are far more difficult to predict than hurricanes and typhoons. And likewise for tsunamis, not every earthquake causes a tsunami. So, one difference is not so much whether it's a lower middle income versus an industrialized country, but just the nature of the disaster itself.

And then a second thing is that prediction and warning systems are typically a lot more developed in the richer countries. So for hurricanes, we have multiple day warnings about what might happen. And then I think a second thing that's important is that if you think about mortality versus property damage, mortality is often higher in more impoverished or poorer settings. The structures are not built the same way. They don't necessarily have engineering standards that they have to adhere to. And so often, it's the case that the death toll is higher in poorer countries than the same event would create mortality-wise in a richer country. But property damage can go the other way because we have, if you think about Miami and the incredible set of buildings right there on the water, dollars of property damage often are higher in richer countries.

Matt Davis:

To what degree do FEMA-like organizations exist around the world? I'm curious, is FEMA an anomaly having something like that or do they have something analogous?

Elizabeth Frankenberg:

No, I don't think FEMA is an anomaly. Indonesia responds to a lot of different disasters because they're on the ring of fire. So earthquakes and volcanoes are pretty common. For Aceh, the tsunami, the disaster was so large scale that they actually built up a FEMA-like organization answering directly to the president. And so that was a little different, and that helped them avoid some of the bureaucracy that tends to exist in organizations that have been around for a long time. But they had a really impressive aid organization and there was a worldwide contribution of funds. So Indonesia was the recipient of something like $7 billion of reconstruction assistance that was truly delivered upon, and it turned into an amazing infrastructure zone in those five years after the tsunami hit.

Lauren Gerlach:

In your studies, I'm wondering what you used to measure exposure, other words, how to identify affected areas?

Duncan Thomas:

So that's a really very good question. And we have, essentially, three different ways to measure exposure that enables us to identify these effects. So the first is, since we have a pre-tsunami baseline that was conducted about 10 months before the tsunami, we know where people were living at the time of the tsunami, and we know their community, where community in this case would be pretty small, maybe part of a zip code would be the way to think about the United States. And we know which communities... Well, when we would start the fieldwork, we didn't know which communities were affected because there was no satellite imagery. And so we would only find out which communities were affected because our teams would arrive at the community and say, "Oh my God, there's absolutely nothing left here." Or people were hardly directly affected.

So what we did was we pulled together satellite imagery, which took a while because it wasn't readily available. And we measured the influx of the water on that. We measured the destruction to built and natural environment. We had some direct observations from our teams, and we also measured max water height. And those provide us with a way to characterize communities as being more or less damaged. And that is a community level measure that has nothing to do with what you were choosing to do prior to the tsunami because as Elizabeth said, which communities were hit hardest by the tsunami was really a function of the topography of the sea floor and to some extent, the location of the earthquake and to some extent, the topography of the land. And since all of our communities are coastal, the probabilities is really quite idiosyncratic.

So that is our first measure. Our second measure is less clearly, completely exogenous, and that is the mortality rate in that community based on our pre tsunami rosters and the number of people who died that we assessed in the tsunami. And that gives us another way to measure a dose response. We've got to think about how much damage there was. But this gives us something about the human life that was lost to the tsunami. So that's our second one.

And then we have a third measure that we use a lot. That is whether you as an individual in your community actually either heard the water, heard the screaming, shouting or saw the water and people struggling in the water, and other similar measures, which are based on our respondent interviews after the tsunami. And those are very salient for these people. And that's three different measures that really are about where you were at the time in the tsunami and how bad that damage was around the tsunami. Those are the three measures that we think when we look at outcomes that are related to those exposures, that we think we can give a causal interpretation to. We've also explored using some post-traumatic stress reactivity indicators. People call it post-traumatic stress diagnoses. We're not doctors, we don't diagnose anything, so we call it post-traumatic stress reactivity. And that's certainly trying to get at a marker of how much stress you as an individual went through and how persistent that was. And that's been very helpful. But that one is much harder to give a causal interpretation to, so we don't.

Matt Davis:

That's interesting. I was actually just wondering, do you even need to measure the storm affected areas? If you're looking at stress, you could just measure loss of income and other types of psychological effects just directly at the person level. It sounds like you're doing a little bit of both.

Duncan Thomas:

We do measure loss of income at the person level. It's difficult to interpret that because of course, if you are not working, you don't have income. And if you are working in a self-employed business with your family, I don't know to whom you will attribute that income. So that is complicated. So then you're going to say, you should use household income, but that's a complicated idea because of course, there are different household structures, household compositions that are going to make interpretation of that change in income very complicated.

And so we have done some work on measures of changes in income, measures in changes in wealth and measures in changes in consumption, actually. And they're all helpful, but they build into them a whole lot of behavioral responses. So, by and large, people who completely lost almost everything in the tsunami, they quickly got back to doing some work of some sort, any sort, because they had to make some income. And one super interesting feature is that a lot of women who were not working pre-tsunami started working post-tsunami because they opened up little food businesses where they would sell food to people and stuff like that. So there's a lot of response around you that you need to take into account in these models.

Elizabeth Frankenberg:

There's also perceptions. And so, of course, people's answers about what happened, how they were affected, embed their own perceptions. And so having the satellite imagery and these models of wave height are an external validation in a way, of what people say about what happened.

Matt Davis:

This question's totally in the weeds as well. I can't help myself because it relates to something I'm working on right now. What did you do about out migration? People moving away, or is that an issue?

Elizabeth Frankenberg:

Oh, yes.

Duncan Thomas:

Oh, yeah. Yeah. No, it's a really, really good question. So, you can imagine after a tsunami that completely devastated the community, a very large fraction of the population would've been displaced. And so when we write the community, there was no one there. And what we did was went to find them. And this is a sense in which this colleague of ours, Cepep Sumantri is just extraordinary. So about half the people who were displaced from the damaged areas went to private homes and other parts of Aceh, mostly Aceh. Some people moved further, but most in the first year after the tsunami was within Aceh. And the other half, approximately half went to camps, most of the camps were military camps that had been repurposed to be places for displaced people to stay. And so we just need to find out which camps these were. We would go there and then do a ton of snowballing until we found our respondents, because we knew their names and their family members and everything, so we could actually find them.

For the ones who moved further away to other places in Aceh, to private homes, that was a lot more work. And that would just be, do you know where such and such went? And then we'd go there and do you know where they are? And so it was a huge amount of work. We have interviewed, of the 25,000 odd original respondents who survived, which is, that's about the number of survivors. It was a little bit more who were in the original pre tsunami baseline, of that 25,000, we have re-interviewed 96% at least once, in the most recent way, which is now 18 years after the tsunami. We have re-interviewed 93% of them, and that is with something like 50% of the people from the damaged communities having moved away. Clearly, if we did not go after them and find them, we would've had a selected sample.

The ones who couldn't afford to move away, or the ones who felt they could do well in this community, you don't even know what bias the selection is. It could go in both directions. And so we are totally 100% committed to following movers. And that was one of the big innovations that we implemented when we did the second wave of the Indonesia Family Life Survey. And we followed movers there and it was really, really hard. And our buddy, Cepep was one of our great supervisors who just found people. It was amazing. He's just really, really, really, very tenacious and very committed to doing the right thing and producing highest quality science.

Elizabeth Frankenberg:

But it was also one of the reasons that working with Statistics Indonesia, to have that pre tsunami baseline and who was in the household, their names, their ages, their relationships to one another, was incredibly important because we knew who to look for as opposed to just, well, who do you think used to live in this household?

Lauren Gerlach:

We're curious what types of cognitive assessments have been done with your study. So I understand you're doing some unique measures including measures of smell. I was hoping you could walk us through some of the different cognitive assessments that are part of the waves.

Duncan Thomas:

So let me focus on the memory-based assessments. I'm happy to talk about the other ones as well. What we're doing is using a smell test to measure olfactory memory. This is work we've been doing with Ralph Lawton, who's a student at Harvard and Mark Albers and Alefiya Albers. Mark is a neurologist at Harvard who designed the assessment and he's used it in a clinical setting and shown that it not only measures olfactory memory, but it's also predictive of AD. So before I describe the test, let me explain why we were super excited to work with these people on this collaboration. In most of these kinds of studies, most of these large-scale population-based surveys, memory is assessed using word and picture-based assessments where you maybe tell a person a story or you give them some words, you ask them to repeat them, and then a little bit later, you ask them to repeat it again and see how good is their memory.

And the difficulty with that is that it is a measure of memory and also, a measure of literacy and education. And it's quite hard to disentangle those in a place where there's very large variation in education. And that is also very, very different across cohorts. So old people have much less education than... Less old people who have much less education than less old people are on the way down that ladder. So for us, the olfactory system was super interesting because first, it's unlikely to be linked to education, and second, it's the most direct way to get to the hippocampus where memory is processed. So the test that Mark designed involves giving people a scent and asking them if they can identify it, and that is a measure of olfactory function. We then follow that up with a recall stage where we ask them, we present them with some of the scents we just presented them with and some novel scents, and we ask them, do they remember having that scent from before. So that's a recall function and that's going to capture both olfactory memory and olfactory function.

And then we do a third stage where we ask them if they can distinguish two scents, which is another measure of olfactory function. And the key to this is that if you just were to do the recall piece, you're not just getting memory, you're getting function, which is related to socioeconomic status and all the rest of it. So what Mark did was he developed a very straightforward method to simply take the recall piece, the function and memory piece, and predicted, using the two function measures from identification and discrimination, and then the piece that's left over is the part that he attributes to memory. So there's a bunch of assumptions built into there, but it's a really clever idea. And so we did that in the field with some sense that seemed reasonable in Indonesia and collaborating with Mark and Alefiya and Ralph.

And then what we showed was that memory declines with age as you would expect, so that was good, but importantly, memory was worse for those people who'd been exposed to the stresses of the tsunami. And that, we know from the earlier comments we made that that's linked to the biological literature to the literature on hippocampal function with stress. And so therefore, we're pretty keen to give that a causal interpretation and provide some evidence in a human population that there are these long-term consequences on memory. When we look at these word-based and picture-based assessments, we see no difference due to stress. So you may think, oh, well, then we don't really believe you that it's really... We like the word-based ones. We have one more thing we do. We added working with Margaret Sheridan, who is a cognitive neuroscientist at UNC. We did an assessment of declarative memory, which is a very different mechanism, and that's the learning and processing and then recall of facts of items.

And so that was done on a tablet where we show people some pictures and they learn what are pairs, and we do that again later. And there we show again, there's deficits among people who had stress, and that's right across the age range. For our scent test, it's really only for people who are a bit older. And so we think that on both of these memory functions, which we believe to be directly related to hippocampal function, we have got a deficit for people who were exposed to the horrors of the tsunami. And the worse the horror, the bigger the deficit. And so we think that's quite likely to be interpreted as causal. And we think maybe these are good markers of memory that people could conduct in other settings where education is... Levels of literacy are not very high and recalling stories is just not something people would normally be doing.

Lauren Gerlach:

I realize your study is ongoing, but what have you found so far specifically related to memory outcomes? I know we've talked a little bit about this, but if you could just tell us a little bit more about what the main findings have been so far.

Duncan Thomas:

So the first result that we have is that the performance of people who are exposed to the worst stresses of the tsunami have worse memory function, both in terms of olfactory memory and also, in terms of declarative memory. We see no evidence of any deficits in any of the other memory markers we have. And we have four of the words survey-based markers that people have used widely. And one of the reasons that we actually decide to try the declarative memory in olfactory memory is that we'd been doing some of these survey-based measures and we never saw any deficits. Even soon after tsunami, nothing ever emerged. And so we felt, we wonder if these memory measures are actually capturing what they thought to capture in the literature? And so that's essentially the main result we have on memory. Plus, we have some evidence that there's no differences by gender, but there are significant differences over age. We see no differences by education, which is great.

Lauren Gerlach:

Interesting. And I know the study's been ongoing for some time, but roughly, how long after exposure do you find that these differences emerge?

Duncan Thomas:

So, we cannot answer that question very well because the declarative memory markers, we first measured 13 years post tsunami, and we've repeated that and repeated and get the same evidence twice. And then the olfactory memory really is a brand new assessment, and we only implemented that actually in the last few months, last six, eight months. And so that would be 17, 18, 19 years after the tsunami. And so would it have been great to have done this in the year after tsunami and two and three and four and traced out that pathway? Absolutely.

So what we can say is they appear to be persistent deficits, and it's unlikely that the deficits are caused by something that’s different from the tsunami, but they may only be emerging later because at your age, memory differences are so small we may not be able to detect them, but as people age, memory starts declining and so we may be able to detect them. So that's a sense in which we feel like there may be premature cognitive aging and our guys are showing these poor memories pretty early on. So we think that there may be some stuff going on and that's why we want to carry on measuring these people. So as they age into the at-risk ages, we'll see if earlier and earlier exposures are showing up in later life.

Elizabeth Frankenberg:

But I think an important thing about the project in general has been that it's not the case that all the bad impacts occur in the first five years and then you recover and move on. There's very consistent evidence in what we see in our data that the effects evolve over time and emerge later after 10, 15 years. And so it's important to think about these exposures as something that have a long reach.

Duncan Thomas:

But it's also important to recognize that there's recovery. So the process is not like a simple linear process. So some people have really showed enormous resilience and recovery and others have not. And we have a whole bunch of results on economic outcomes. And briefly, what we find is that people's economic status has really improved dramatically after all the horror of losing everything they owned, but they haven't actually caught up with their peers who didn't lose everything. So their economic status has improved faster than the others, but not to the level... They haven't reached that same level. So the stories are extraordinarily complex and not ones where there's a simple storyline that everything is terrible or nothing is terrible, but rather... And they vary by domain.

Lauren Gerlach:

I was curious and speaking to that, if you found that there are some factors that help mitigate these stressors or across any of the studies you found, some things that are linked with that resiliency or folks who tend to cope a little bit better associated with less cognitive decline? Anything that you've started to see a signal with that?

Duncan Thomas:

So, I think that there can be no doubt that connections with family are very, very important or connections with family and friends are very, very important. And I think engagement in the society is very important. We have one set of results, which I think are quite stunning, that is that quite a few kids were orphaned by the tsunami. Orphaned or at least lost one parent. Some lost both parents. Now, not many kids under the age of say eight, lost a parent because if they were that young, the parents would've saved them. And if the parents died, everyone in the family died. But with kids aged eight to say 17, we see evidence, we can trace out what happened to them. And the thing that's really shocking for us when we looked at this was the eight to 15 year olds, we see no impact on their education levels and we see no impact on any other markers that we've looked at in terms of economic outcomes.

But for the kids who were 17, 16, 17, 18 at the time of tsunami and lost a parent, they left school early and the boys went to work and the girls got married. And so that may have long-term consequences. And so I think the implication there is that the family, the extended family, the community, really came together to try to help the kids they could help. Whereas the older kids were looking to look after their younger siblings.

And we see some other evidence, for example, that a lot of people lost a baby and so many of those people went and had another baby afterwards to rebuild their families. But a large number of women who'd never had a child in communities that lost lots of young kids, they had kids earlier than in communities where there was no death. And Elizabeth conducted some qualitative work and they said, "We're trying to rebuild our community." So I think that family, community was really important. And there's no question, economic resources, the phenomenal reconstruction program post tsunami, all of these led to enormous resilience and recovery. But the individual characteristics is an incredibly complex story that depends on which domain you're looking at.

Matt Davis:

So as you know, there are a lot of different risk factors when we talk about dementia, and it's likely probably hard to unpack, but I'm curious if you have any sense for the effects specifically of stress compared to other known risk factors when it comes to dementia?

Duncan Thomas:

So I think that's really difficult to pin down. We've documented the causal impact on memory. We also have some evidence on HPA axis activation. Specifically, we see that women who are exposed to the worst of the tsunami, 15 years later, they present with reduced cortisol and that's interpreted as burnout. And we know that that has linked to worse health outcomes. We know that elevated stress has resulted in higher CVD risks. There's elevated inflammation, greater adiposity in the population, but we also see some positive effects. So for example, we see that people who were exposed to the worst tsunami have better executive function. So we think that's likely because they had to navigate a new landscape to remake their lives. I don't know that we could say that there is a causal effect on dementia because all of these factors are going to play into dementia, and some are negative, but some are also positive. And so how are they going to play out is going to be a function of which ones are going to dominate in that person's life.

Matt Davis:

There's probably a lot of interactions too, I imagine between.

Duncan Thomas:

Absolutely.

Matt Davis:

Yeah.

Duncan Thomas:

Absolutely. Which is why I think having these multi-dimensional studies where you have many domains of health, cognition, how these people behave, their activities of daily living and their connectedness with family members seems so important to me and tracing it over time.

Lauren Gerlach:

As a follow-up, at the population level, to what degree do you think the effects of the earthquake have contributed to the region's dementia prevalence?

Elizabeth Frankenberg:

So I don't think we know the answer to that yet, and we may never totally know that, but certainly, it's clear that there are these long-term implications for physical health and aspects of physical health that seem to be risk factors for dementia as well. So it won't surprise me if it turns out to be the case that one can make that linkage, but I don't think we're at the point where we can say that yet.

Matt Davis:

So in terms of your team's work, what's next?

Elizabeth Frankenberg:

Well, we'd like to do another round of STAR and capture the 20-plus year experiences that people have had. That'll be a point where we can start to look at, for the kids who are the smallest at the time the tsunami came ashore, how their education and family formation and entry into the labor market are starting to really play out. And we'll also be able to look and see whether there's intergenerational aspects of transmission of trauma in terms of kids born after the tsunami to badly affected parents, show any impacts. And then as we look at what's happening to the older aging population, we're going to start to see the benefits of having started tracking memory and cognition five or 10 years ago on long-term implications. So hopefully, we'll get another wave in the field because there are all these puzzles that we want to keep trying to unravel with the data.

Matt Davis:

That's exciting to hear because I was thinking, I was like, you are set up perfectly for a birth cohort type study.

Duncan Thomas:

And in fact, that's a very interesting group to... You're absolutely right. A very interesting group to study would be the people who were in utero during the tsunami and the kids who, their older siblings, and the kids who were born afterwards in this repopulating. And so we've done some work on the in utero population and seeing how do they play out when they're 20 and 25 would seem really, really, really important. And we're seeing evidence of the kids who are exposed in age 10 to 15, for example, are presenting with very high rates of hypertension, like very high rates of hypertension. So what does that portend for the future of CVD? There's lots and lots of really, really important questions here.

Lauren Gerlach:

Anything that we haven't covered that you'd like our listeners to know?

Duncan Thomas:

No, I think you've done a great job. I think the only thing I would say is that it's really from our perspective, how to overstate the importance of conducting these long-term longitudinal studies. Because a lot of studies would say, go and find people 20 years after or 40 years after some disaster or some event, the 1918 Spanish flu or some earthquake somewhere. And what they'll do is they'll link people back to where they were at birth, but they don't know what that person's exposure was. They just know that they might've been in that area.

And what we have is really tight measures of exposure that we talked about, and then we can follow out and say what kind of stresses these people were exposed to and what resources they had around them. And then say, "Who has shown resilience and who has really never recovered?" And that's the only way you can do that is with these kinds of long-term, longitudinal studies. And as long as we continue to follow very large fractions of the 93, 95% of the original sample, it feels like we can actually contribute to science in ways that would speak to important questions about, for example, climate change.

Elizabeth Frankenberg:

And then just for you guys, if you're interested, there's a study on the front page of the Washington Post today, I think, about the risks that the West coast faces from a major tsunami and the unpredictability of when it's going to occur.

Matt Davis:

Well, this has been great and super interesting. Thank you so much for joining us.

Elizabeth Frankenberg:

Absolutely. Thanks so much for having us.

Matt Davis:

If you enjoyed our discussion today, please consider subscribing to our podcast. Other episodes can be found on Apple Podcasts, Spotify, and SoundCloud, as well as directly from us at capra.med.umich.edu, where a full transcript of this episode is also available. On our website, you'll also find links to other resources we've created specifically for dementia research.

Music and engineering for this podcast was provided by Dan Langa. More information is available at www.danlanga.com. Minding Memory is part of the Michigan Medicine podcast network. Find more shows at michiganmedicine.org/podcasts. Support for this podcast comes from the National Institute on Aging at the National Institutes of Health, as well as the Institute for Healthcare Policy and Innovation at the University of Michigan. The views expressed in this podcast do not necessarily represent the views of the NIH or the University of Michigan. Thanks for joining us, and we'll be back soon.