If you’re new to dementia research, you’ll soon come to find that most research papers on dementia start off something like this: “In the United States there are 5.8 million individuals living with dementia and this is expected to increase to 13 million by 2015. . . .” In this episode we discuss the study on dementia prevalence that has been cited thousands of times with one of the authors. Dr. Jennifer Weuve from Boston University joins us today. We also talk more broadly about what makes a research paper highly citable in general.
Related Links
CAPRA Website: http://capra.med.umich.edu/
Transcript
Matt Davis:
If you're new to studying Alzheimer's disease or just starting to read studies about the condition, you'll soon come to find that the majority of articles start out something like this. In the United States, 5.8 million older adults suffer from Alzheimer's disease, and the number is expected to grow to 14 million by 2050. Dramatic reading, very much intended. Then the article will go on to talk about why their study is so important, or if it's in a grant proposal, why their work should be funded. The study where these statistics come from was published in 2013 in the Journal Neurology.
At the time of recording this podcast, the paper had been cited over 2,400 times in other research articles. For researchers, the number of citations is analogous to performance statistics in a sports competition. Say for a running back in football, it's like the total number of yards in a specific game. In this case, the study being the game, but I digress. The metric score that's a composite measure of media attention for this article is over 600. That places it in the top 5% of all scientific articles, and it's been the focus of more than 60 news articles.
Today, we're going to dig into the paper that is most often cited in the first sentence of Alzheimer's studies. We'll get into what exactly the researchers did, the strengths and limitations, discuss what makes some articles so highly cited and others not so much. I'm Matt Davis.
Donovan Maust:
I'm Donovan Maust.
Matt Davis:
You're listening to Minding Memory. Today we're joined by Dr. Jennifer Weuve. Dr. Weuve is an Associate Professor in the Department of Epidemiology at the Boston University's School of Public Health. She's an epidemiologist whose research focuses on two areas of inquiry. One, factors related to aging and two, health effects of exposure to environmental toxicants. Dr. Weuve has an impressive track record as an investigator on numerous prestigious grants from the National Institutes of Health.
Small world, Dr. Weuve was the first author on the paper we discussed with Sara Adar on this podcast earlier in the season, the one on community noise and cognitive function. She's here today though to talk with us about her work on what is currently one of the most cited papers on Alzheimer's disease. Jennifer, thanks so much for joining us.
Dr. Jennifer Weuve:
Thank you for having me.
Matt Davis:
Dr. Weuve is the second author on the paper titled Alzheimer disease in the United States. (2010-2050) estimated using the 2010 census. Being an epidemiologist, we thought Jennifer would be the perfect person to talk about the details of the study, further, given her success, a great person to help us try to answer, what makes a paper citable? Before we jump into the discussion, just a little bit about the study. The study set out to make prevalence estimates for Alzheimer's disease. Overall, I would say it's a straightforward read.
After seeing so many paper cited, I was actually surprised that it's fairly short in length. The authors used a combination of data sources to forecast the number of older adults with Alzheimer's disease from 2010 out to 2050. The data source used included the Chicago Health and Aging Project and the 2010 US Census. Just to start things off, Jennifer, when you were working on this study, did you in all anticipate the effect it would have on the research community?
Dr. Jennifer Weuve:
Yes and no. Here, we were going to go about trying to identify a cause of dementia or a treatment for it. Really, that's what people are very interested in because this is such a devastating condition. We were aiming for something more fundamental. How many people have Alzheimer's dementia now? Back then, now was 2010, which seems like another world. How many will have it in the future? This seems like a fairly basic aspect of dementia. I mean, it doesn't seem as enthralling as an investigation of say a new effective treatment for Alzheimer's disease or some highly effective way to prevent it.
Yet, of course, as you were getting at in your introduction, knowing how many people have Alzheimer's dementia or at least knowing approximately how many do help us a scientist also as funders of science to prioritize what we have researched. It also helps us as a society plan. It's hard to argue that Alzheimer's disease merits a certain kind of research attention if you don't really know how many people have it, if you don't even have a sense of it. We already know it's a devastating condition though. It's also hard to make decisions about how to allocate resources to people who have dementia and the people who care for them.
By caring for them, I mean family members, but also professional caregivers, like physicians and nurses and PAs and so on. When we started this study about 10 years ago, the most cited estimates of Alzheimer's dementia prevalence were from a paper that was published in 2003. This paper projected the experience of a specific well-studied cohort onto the US population. That study was based on census data from 2000, and so that meant for example that the expected number of older adults in the United States with Alzheimer's dementia in the year 2030 was based on data from at least three decades prior.
That's old. If our prediction game is good, it shouldn't matter, but it just starts to feel a little wobbly the further into the future we go. So, that in essence was the goal. I realize now in hindsight, I wasn't that naive about the importance of providing a citation for the first sentence of many papers, but in fact-
Matt Davis:
It wasn't your goal.
Dr. Jennifer Weuve:
... it was not a goal. Every house has a foundation. Every home has a foundation, and essentially, this is the foundation for many papers. And so, it is obviously a privilege to have this work sided so often and to have it viewed with the respect that it's received. In fact, my role as the second author in fact was not to conduct the analysis, actually. It was to write the paper. And so, in that position, I became even more intimate with the procedures for generating these estimates. When you're doing it, of course you know it.
But when you have to explain it, that is another level. So, it is definitely a nice match or a nice concordance between the work and care that went into this and the attention it's received. I know that's not true for many papers.
Matt Davis:
There is beauty in doing something that needs to be done well, I think. Sometimes scientists think that they've got to come up with some super novel association or something, but sometimes we just need the numbers, and that's one of the things I really appreciate this article.
Donovan Maust:
That was a great segue into the next question. On the surface, it seems like coming up with prevalence estimates would be a pretty straightforward thing to do and yet if you read the methods of this paper, it's not so straightforward. And so, you just were explaining that you were the person who really had to write this. Can you just explain, give us an overview, of some of the nuts and bolts of how you actually came up with these numbers?
Dr. Jennifer Weuve:
Yeah, okay. First, let's just have a little recap of what we mean by prevalence. Prevalence, it's kind of two things, but they're very related to each other. We talk about prevalent cases or prevalent dementia as the number of people who have Alzheimer's disease at a given time. Usually, for Alzheimer's dementia, we speak of the number of cases on average for a specific year. And so, if we say the prevalence is 5.1 million, what we were saying is 5.1 million people have Alzheimer's dementia for say 2010.
We can also talk about prevalence in terms of proportions, and basically all we're doing there is we're taking that first number that I described, the prevalent number of cases, and dividing it by the total population. For example, if we are interested in the prevalence proportion of Alzheimer's dementia in say the 85 plus population, we take the number of cases and divide it by the number of people who are 85 or older. Both of those things have value, and as you can tell, they're very closely related.
Let's start about how many... Well, actually, let's start where many people start, which is, why not just count up the number of people whose medical records show that they have a diagnosis of Alzheimer's dementia? Why don't we just do that? It can't be that complicated. This is similar to the approaches that we use for conditions like some cancers or even a neurologic condition like ALS. With conditions that are severe or seem to develop suddenly are definitely unusual or even conditions for which we screen people on a regular basis, most of these folks end up at the doctor's office.
With Alzheimer's dementia, the situation is quite different. Historically, dementia was believed to be a normal part of aging. Oh, just having a little deficit there. Well, you're old. What did you expect? On top of that, there was and still is stigma around having this diagnosis. Finally, even though there are ways that folks with dementia can manage and adapt to their symptoms, there's no treatment that truly alters the course of the clinical disease. So, all of these factors plus ones I haven't even mentioned mean that Alzheimer's dementia is often not diagnosed. When it is diagnosed, it may have been around for quite a while.
And so, if we use medical records, we're going to get the wrong number, and this is even in societies where everyone has medical insurance. Technically, in the US, everyone's 65 and older has it too, but it doesn't really seem to matter. In terms of how far off we might be, if we use claims, insurance claims, or if we use medical records, maybe 15%, but maybe more. It just depends. We're just not going to get it right. One way around this is to take the diagnoses into our own hands. And so, once we worked with a sample of older adults and evaluated them ourselves, this evaluation would be the same for everyone.
We're not going to depend on everyone, all physicians in the community diagnosing people in the same way. We're going to do it ourselves, or we're going to do it the same way for everyone, and that means it won't depend on having a doctor or going to one. It would also occur again in our fantasy situation on a regular basis, so we could identify people who have Alzheimer's dementia and people who newly develop it. If we're regularly evaluating people, we can catch them as they're developing it. That means that we're not going to have so many diagnoses that are late in the course of the disease.
Of course this attention to the consistency of the diagnostic process and the regularity of evaluation is why we can't evaluate the entire population of US adults. We just do not have the funds for that. So, to get around that, we could sample older adults. And knowing things about our sample, we could then use that information to project to the US population, and that's what we did. Well, in our case, we used participants from the Chicago Health and Aging Project, and we used information on them to project to the United States.
Donovan Maust:
One quick follow up is, the title of the paper specifically is Alzheimer's dementia.
Dr. Jennifer Weuve:
Mm-hmm (affirmative).
Donovan Maust:
How specific is this estimate in the clinical exam really is Alzheimer's dementia? And so, other types of dementia we've heard about earlier, like dementia with Lewy bodies, your frontotemporal dementia, those folks are not included in this estimate?
Dr. Jennifer Weuve:
Yeah. I'm really glad you asked that question. This study used a definition of Alzheimer's dementia that could be considered Alzheimer's plus. Anyone who met the clinical criteria for dementia but also had symptoms of other dementias was also counted as an Alzheimer's dementia case. This was an insight of the principal investigator, Dr. [Denis 00:13:29] Evans, who came up with this insight a long time ago before we had a lot of pathologic data to support it. That sort of approach has been supported now by what we see on pathology.
In fact, if you counted all the people who have dementia in this particular cohort, the dementias that are clearly not at all Alzheimer's, at least [inaudible 00:13:55] clinical presentation, account for like 5% or 6% of all dementia cases. What we're probably seeing, and I'm sorry, I don't actually have these numbers on me, is a lot of mixed dementias and then what some people might consider pure Alzheimer's. But the longer I've been in this career, the less I believe in it. [inaudible 00:14:18] Alzheimer's or it's not as common as we think.
Donovan Maust:
Right. So this is pretty... This would capture the vast majority of dementias, perhaps, except for a very small, small percentage?
Dr. Jennifer Weuve:
Yes.
Matt Davis:
I was going to ask whether you'd feel comfortable in that first sentence being dementia in the United States, not just Alzheimer's dementia.
Dr. Jennifer Weuve:
Yeah.
Matt Davis:
It sounds like you [crosstalk 00:14:39]-
Dr. Jennifer Weuve:
I would actually prefer just to say that, but it's sort of, I'm adhering to the terminology that we used. And so, thank you again for asking that. It's something I'd forgotten to take note of, and it was going to be one of those questions I hope you asked me and you did.
Matt Davis:
So it sounds like you were taking into consideration people developing the condition and then mapping that to the US data. What about mortality data?
Dr. Jennifer Weuve:
Yeah, okay. Let's talk about prevalence again. The number of people who have any condition of interest, and then let's just again stick with all the of dementia, it's going to depend on two things. Think of it as, where are the cases coming from? Well, they come from newly developed cases, but there's also I guess an outbox. It's not really an outbox-
Matt Davis:
[crosstalk 00:15:28].
Dr. Jennifer Weuve:
... but there's also cases leaving the population and those are people who die. And so, in fact, what we estimated in the Chicago Health and Aging Project, which I will now call CHAP, what we estimated were two things, two major quantities. One was the incidence of Alzheimer's dementia, and the other was the relative rate of mortality among people who have Alzheimer's dementia. And so, by relative, how much more rapidly did they die compared with people who didn't have a condition?
Both of those two quantities, we estimated combinations of four variables: age, sex, race, meaning black or white, and education. We had unique incidence rates and relative mortality rates for every combination of those four variables, and it was through that combination of incidence and relative mortality that we estimated prevalence.
Matt Davis:
I was having flashbacks to my MPH degree because we used this textbook by Leon Gordis. Have you seen this book? It's a famous textbook.
Dr. Jennifer Weuve:
Yes, I am... Yes.
Matt Davis:
Yeah. There's an image that he has of this flask to describe prevalence, and each little ball in there is a person. He has people developing the disease and then potentially leaving the population and [crosstalk 00:16:56].
Dr. Jennifer Weuve:
Leaving the prevalence pool.
Matt Davis:
Yeah, exactly.
Dr. Jennifer Weuve:
It's funny because I was going to say prevalence pool, and it just sounds so jovial compared with dementia itself and just didn't feel like in the same tone family.
Matt Davis:
You touched on this already. But in terms of the CHAP study, the Chicago Health and Aging Project, was there other things that we should know about that particular study in terms of the quality of the data or why you selected it?
Dr. Jennifer Weuve:
Yeah. Yeah. The Chicago Health and Aging Project involves nearly 11,000 older adults. By older adult, I mean 65 and older living in four adjacent neighborhoods in Chicago. Every three years, starting in 1993, participants underwent evaluations of their cognitive function and so on in their homes. So, you can see how we are really aiming for the population experience. We are not asking anyone to come to the clinic. We are coming to them. The other thing you should know about this population is that about 60% of the participants were black.
The remaining participants were white, and participants were recruited into the study over time. So, for people who study children or even younger people, we talk about the equivalent of being born as being newly aged, which means you turned 65 and you were suddenly eligible to be in this study. And so, in fact over time, the investigators recruited more people into the study as they turned 65 so long as they lived in this neighborhood.
Donovan Maust:
In the analysis, you stratify or I guess stratify the analysis by age group, so you have 65 to 74, 75 to 84, 85 plus. When in figure one in this paper, the overall line is going up and then you see these interesting shifts over time within those different age strata, which has something to do with the aging of the baby boomer cohort. I was wondering if you could just tell us a little bit about the overlap of risk of dementia by age, and then what happens as the baby boomers are aging to help explain what we're seeing with those different age groups?
Dr. Jennifer Weuve:
Yeah, sure. Age is a part two aspects of our results. One which is going to be really familiar and which you touched on, let's start with the percentage of people who have Alzheimer's dementia. And so, because age is such a strong force in determining risk, we see risk increasing exponentially with older age. The older the age group, the larger the percentage who have Alzheimer's dementia. That, we saw very, very clearly in our data. For example, for 2020, we predicted that about 3% of those 65 to 74 have Alzheimer's dementia as opposed to nearly 17% of those in the next age group, 75 to 84, and then 32% of those 85 and older. No surprises there.
That's the one way age works. Basically, the older you are, the higher your risk for having dementia. There's mortality elements there too. I'm not going to get into it at the moment. Now, I will say these percentages moved around a little bit from year to year, and that starts to get it at this cohort effect. It's not a lot. For example, you'll see that the percentage of 65 to 70 year olds moves up and down over time. Part of that is that, okay, now I'm going to start to sound really nerdy, the composition of the population within that age group changes too. That's partially because of the aging of the baby boomers.
For example, 65 to 74 year olds, over time as we get to about 2030, that group becomes older. They're still 65 to 74, but there are more people toward the 74 end. So, we see. Those ships that we see in the numbers over time are because of the composition within aging cohort. Again, that is really fed by this aging of the baby boomer cohort. And then the aging of the baby boomer cohort also affects the trends that you mentioned, the trends in the number of people who have Alzheimer's disease dementia.
And so, surely, I think the magic year is about 2040 is in there. Surely, before 2040, we projected that the number of 75 to 84 year olds with Alzheimer's dementia would level off. It's going up, and then it levels off. But the number of 85 plus year olds with Alzheimer's dementia, we projected it would continue to grow. This is mainly because of the continued growth in the 85 plus population, and that is the baby boomer population. That's the bulk of it.
So, really, before 2040, we expected most persons with Alzheimer's disease to be 75 to 84. After 2040, we expect most of them to be 85 plus. And so, this image I have in my head, which again just seems inappropriate, is like a snake that has swallowed a very large animal and the animal [inaudible 00:22:23]. And so, we are seeing this [crosstalk 00:22:27]-
Matt Davis:
Those are the baby boomers.
Dr. Jennifer Weuve:
... swallowed by a snake. Yes.
Matt Davis:
For any epidemiologists that are training, this is a great example. Can you call this a birth cohort effect potentially? If you do-
Dr. Jennifer Weuve:
It is totally a birth cohort effect. Yes.
Matt Davis:
That's something that's so great to hear in reality because often, the examples in textbooks are like tuberculosis and strange things that we don't really talk about anymore. It's funny. I'm so glad that you mentioned that because when I first looked at the study a couple of years ago, I didn't think that much about it. I just looked at it quickly and I was like, "Why are they stratifying?" Until I read it more carefully preparing for this interview, I was like, "Oh, that's exactly why they had to stratify these results." Because it would be this artifact if you didn't do that. So, that was a great explanation.
Dr. Jennifer Weuve:
Yeah, it is. It is definitely a birth cohort effect. I think I also am not like a spring chicken here, and I realize I have to now explain to many of my students what the baby boomers are. This is like war and then people came back. Anyway, I don't know if you'll need to add an explanation onto this for any of the earlier career people.
Matt Davis:
The focus of your article seems to be on counting up the number of people and projecting that. But when we think about prevalence as a proportion or a percent of the population for something that's fairly common in older adults like dementia, do you think that the prevalence as a percent is increasing in the United States?
Dr. Jennifer Weuve:
Yeah, that's a really good question. I should have actually thought about this more. Let me... I'm going to think aloud. It's going to be ugly. Is the prevalence as a percent, do you mean prevalence as a percentage of the entire population?
Matt Davis:
As a percent of older adults, I would say. You do look at the percent by age category and they do seem to go up across it, but it's hard to unpack that a little bit because it implies why... I started asking myself, is it increasing or not? Why would it? Are we diagnosing it more or recognizing it more?
Dr. Jennifer Weuve:
Oh, Oh, okay, okay. Yeah, okay. These are different things. All right. First of all, there's dementia in the abstract and then there's dementia in what we measure. Let's assume that anything that comes out of CHAP and is projected to the US population or not, that we assume that because we have more control over the measurement process, that there's less of a measurement artifact. That we're picking up everyone who has it and we're not identifying anyone who doesn't as not having it. What did I say? As having... The detects are true, truly had dementia, and the people who we don't think have it actually don't.
So, the prevalence in the US population is going to go up so long as there are major at risk populations that are increasing. Imagine we have... We can take the baby boomers, for example. They are currently entering... They are entering their mid 70s on the high end. The earliest baby boomers are entering... They're like 75 ish now. And so, I'm just going to pant here. I think this is going to keep going up. The prevalence is going to keep going up, but then what's going to happen is that there will be fewer and fewer baby boomers, and it's going to take a while, by a virtue of having died.
As the baby boomers die, this sounds horrible, the population of older adults is just going to be smaller. And so, we will actually not have as many people or we won't have as high of a prevalence overall, but it really is going to depend on what the age composition is in the older adult population.
Donovan Maust:
This might be wrong and also might be that I'm being messy with my terminology, but my understanding, I thought there was a paper from HRS, from our co investigator, Dr. Langa and then I think also one from the Framingham study that suggested the observed prevalence of dementia is in fact slightly lower than might have been anticipated a decade ago. That's on a absolute number that's bigger, so it's more people, but the actual proportion percentage is a little bit lower than what we were expecting, right?
Dr. Jennifer Weuve:
Okay, right. Okay. The assumptions in our projection is that anybody's risk of developing dementia in any given year is not going to change. If you are a 75 year old say black woman with a college education, we are going to assume over time that you are going to... That any woman who has those characteristics is going to have the same risk, no matter, from year to year. So, what this-
Donovan Maust:
Just to be clear, you mean a person with those characteristics in 2010 versus a person with those same characteristics in 2030, the risk will be the same? That's what you mean?
Dr. Jennifer Weuve:
Yes, exactly. Exactly. We would say... For example, let's suppose the risk is say 8%, we assume it's going to be 8% no matter what, no matter what year it is.
Donovan Maust:
For a person with [crosstalk 00:28:01]?
Dr. Jennifer Weuve:
That will have been based on something from even earlier. What the Framingham HRS studies did is they actually directly observed people over time. We are projecting in the future based on the present, assuming that the present going to keep operating the same way in the future. What they did is they just said, "We're just keep looking at people over time, and we're going to estimate prevalences in that are, I guess, contemporaneous to our data." So then what they have suggested is that maybe this risk is going down, but as you say, the number of people is still going to go up because we have a larger, a growing number, of people who are 65 plus.
That is going to just drive the absolute numbers even if the risks start going down. We actually tried this in CHAP too. We tried looking at trends over time, and we had to stop at around 2010 because that's as far as our data could take us. We didn't see the same obvious pattern. I think that also begs the question about, well, first of all, if these, I guess, improvements in risk or incidence are real, and they may be, and how universal they are. Most of the studies that look at these things involve white people, and that goes for studies in Europe as well.
CHAP is a little different in that regard, and we just didn't really see that strong a pattern. There may be something, but it's not screaming. So, this is definitely a place where we could do a better job, for sure.
Matt Davis:
This might be a little more granular than we want to get into too much, but in your paper, you did mention several different sub-analysis that you did to double check whether your prevalence estimates would change based on different parameters and assumptions. I'm just curious, overall, did those analyses make you more or less confident in your estimates?
Dr. Jennifer Weuve:
Well, overall, they made me more confident. I wasn't around for the first paper. Most of the team actually was. And so, it's a new kid block. The view of the team overall is that these are the... They're very educated guesses. They are quantitatively educated guesses. And so, their view of what a precise estimate is is very... I found it to be quite different than what a lot of the consuming public views. The consuming public sees the 5.1 million, they're like, it must be 5.1 million. Having worked pretty in-depth with this team, I can actually picture now how they see this. They see a like 5.1 dot and space with all this cloud around it.
And so, anyway, I think I would like to echo their view. They think this is pretty consistent. Essentially, what happened is that we used the same source data essentially with some modifications. The 2003 paper involved CHAP data that ran from about 1997 to... I'm going to forget this now. I think it's 1997 to 2000. And so, not lot of data. They also relied on the US Census from 2000 as well. Again, this is pretty old data projecting the future and the data on which the incidence and relative mortality was based was pretty limited.
It's four years, essentially. 10 years later, we have more data from CHAP. We also have a new census, so that means that say a population projection for 2020 is not going to involve so much guesswork as it did in say the year 2000. It's kind of shocking then how similar these two estimates were given that our estimates are based on other estimates. Where others may see a kind of draft or it went up, in fact, I'm actually, they're pretty consistent actually. To get into the weeds on this, the additional data that we had for CHAP, we had an additional 10 years of data pretty much, that changed the 2050 projection more than it did the 2010 estimate.
It wasn't a lot, but it changed a little bit. Updates to the data we had on the United States population, those changed the 2010 and 2050 estimates a little bit. They changed down. Actually, the education data shifted those estimates downward a little bit, which suggests to me that in fact maybe the changes in the census data, such as more people had more education, that's what I would expect with that. And then the biggest change actually resulted in an upward bump to the 2050 projection, and that was from just overall changes to the population of itself. I'm not sure what specific changes there were in that projection, but that is not incredibly surprising. Again, projecting out to 2050, it is really shocking how similar these are.
Donovan Maust:
We've been talking about the strengths of your great paper, and the 2,400 citations suggest that there are lots of strengths. If you had to say what you thought some of the main limitations were, what would you say?
Dr. Jennifer Weuve:
Yeah, I like to wear the limitations of my papers on my sleeve. Mainly, I feel like these are things we should talk about openly because we should be promoting advances in science and understanding and so on. I'm just going to go back to one of the hallmarks of this study, which is that it was really based in estimating incidence and relative mortality in groups defined by four characteristics: age, sex, race, and education level. These estimates did not account for anything else in the population that might affect incidence or mortality. Age is a big player. It is pretty much the player.
We think about some of these autosomal dominant genetic variants. Sure, they have a very strong effect on risk, but they're so rare that they would make very little difference in this kind of estimate. So, what other factors might we be interested in? As I was thinking through this and I was thinking, well, okay, let's take physical activity as an example. Physical activity does seem to affect mortality risk. It probably reduces it, and it probably reduces dementia risk, we think.
So long as the physical activity patterns in CHAP are like those in the United States and if they stay the same over time, then probably this very simple approach that we took is not a problem. But if CHAP is really different from the population at large, our estimates could be way off. And so, related to that, our estimates accounted for the growing number of black, older adults in the United States. I mean, so many of our participants were black, but we didn't account for the emerging trend in the older adult population that entails the growth in the number of people who are Latinx or Hispanic, people who are Asian or some combination of ethnicity in race.
These populations are growing quite rapidly. Recent studies do suggest that the risks in these groups are different than the risks of those who are white or black. Even within subgroups of these populations, we see differences in risks. So, we did foresee this limitation, but we couldn't really do anything about it. There were very, very few his Hispanic people in CHAP. Again, I think maybe I view that as a weakness, but also as a message as to where this estimate should go in the future and how maybe these studies should be conducted in the future. We did foresee that limitation.
What we didn't anticipate, and I don't know yet how big a limitation this is, but we didn't anticipate the emergence of COVID-19, a global pandemic that has shortened the lives of more than half a million older adults and may have placed many others at unknown neurologic risk. We don't have a firm grasp yet on how COVID-19 has altered that landscape. Part of me thinks it's going to be a blip, but I don't know. I think the jury's out. And so, hopefully that would be something that also gets absorbed into future research.
Matt Davis:
That is a great overview of the study. I learned a lot about some of your thoughts into the paper that went into some of the decisions that you made. Just to come full circle, I'm curious if we could start to wrap things up by just talking a little about what we think. As several researchers here on the line, what drives some articles to be more highly cited than others? Because I think that is an important thing when we write and perform analyses that we're doing things that have value to other researchers and that can have that kind of impact.
Dr. Jennifer Weuve:
Yeah. Well, I think this idea that papers take on a certain structure and sub structures. We have the intro, the methods, the results, and the discussion. But even within the intro, usually it's like, what is the problem that I'm addressing? Or essentially, what is this thing? And why is it important? And why should we care? In a grant, it's even more of a marketing statement. It's like, this is a terrible, terrible problem, and you must give me money to study it.
And so, I think there's a kind of paper that forms the foundation of numerous paper that certainly leads to higher citations. Other kinds of papers that make... Well, yeah. I'm thinking back [inaudible 00:38:09], but now I'm trying to think of some other highly cited paper. I was also involved with a group also at my institution, my current institution, BU School of Public Health and the BU Medical Center, and this is a group that studies a chronic traumatic encephalopathy. They developed some diagnostic criteria.
I am part of that paper. I'm way in the middle, so I don't claim to speak as much in-depth about it as this one, but that has received a lot of citations too. But we can think about this in the same way. This has become the main ingredient to a lot of papers that have followed. If you're going to study CTE or sports injuries involving head trauma, you're probably going to cite this thing, especially if you're using diagnostic criteria that are now widely accepted. Back to one of your observations, you said, often people like the boring parts. These are boring elements of research [inaudible 00:39:15].
We don't discover something. But in fact, we kind of did. We discovered how many people have dementia. I'm going to go on the slight tangent here. The PI of the study again, Dr. Denis Evans, was one of the first people to conduct a major epidemiologic survey of dementia in the population. This was conducted in East Boston. At the time, the knowledge about who develops dementia and how common it is was pretty limited. One of the assumptions that one of my dear mentors has shared with me was that, for a while, some people thought that higher education was associated with higher risk, because guess who was coming to the clinic?
And so what Dr. Evans and his colleagues discovered when they went door to door in East Boston is they discovered more people had dementia than we ever thought possible. This was a really, really big deal. He later told me, he said... We were actually going to be, I think it was on 60 Minutes and then the Berlin Wall fell. And so, anyway, I just wanted to share that little tidbit.
Donovan Maust:
For biomedical papers at least, there's a standard three paragraph introduction, and the first paragraph is always the problem. Especially if you're going for a general medical journal, you're trying to convince cardiologists or surgeons or whoever what the problem is, and the way you do it is with numbers. And so, literally, this is like the first sentence of the first paragraph of the paper. I also wonder too a little bit about the timing. You mentioned in your introduction, which I'll note is also amazing, the one paragraph introduction in this paper-
Matt Davis:
I love it. I love [crosstalk 00:41:10] intro.
Donovan Maust:
... Which is so amazing, but you referenced the National Alzheimer's Plan in the US. And so now, there's been this, what, 10-year period of sustained investment and attention with I think a lot, maybe more, people coming into this research space. So then you all were like, "Oh, look, here's this paper that all of you new people coming in can cite in your grants and papers and projects that you're working on." It seems like the timing for this particular analysis was also just perfect given the national interest at the point.
Dr. Jennifer Weuve:
Yeah. Well, historically, there wasn't a lot of funding put behind Alzheimer's research. It was a pretty dark place to be, or at least I would say inconsistent. There would be great periods followed by pretty dry periods. It is these statistics, and again, this is through many layers of translation, they end up on the desks of say senatorial aids or something. These are lobbying numbers as well. We use it to justify the importance of our research, but it's also used to justify why this research should be funded, why the fund should allocated. And thus, over the past few years, we've never seen anything like this in terms of funding that is available to study Alzheimer's dementia.
Matt Davis:
It's so good that you focused on prevalent cases though. My questions about the proportion were one thing, but that's what you need. When you think about resource allocation for something like this, that's exactly what everybody needs. That's I think why so many people cite this, is that the sheer number of people we have to consider as we move forward as a society. Something else I'll just throw out there in terms of citable papers is, I think that you do a really nice job of minimizing jargon and presenting tables and figures.
I feel like epidemiologists have a great appreciation for clarity and simplicity and their tables and figure design, and I feel like it shines through in articles like this. There's a fair number of methods. Once you start reading the methods, you say, "There's a lot of detail and some of the inputs and outputs and everything," but the presentation results are something everybody can wrap their head around and everybody can get and everybody can make sense of, and I really appreciate that about it.
Dr. Jennifer Weuve:
Oh, thank you so much. Well, I hope that most of the work that I'm involved in or at least I'm in charge is useful. Useful is ultimately the goal here. Even though the cynical side of me says, I'll just throw out any garbage and just get your citation number up or add a new page to your CV and get promoted or whatever. But I think the satisfaction of being in this field, being a scientist in general, is knowing that people are reading and using your work and even building. Well, specially building off of it and improving it. I'm all for that.
Matt Davis:
Jennifer, thanks so much for joining us. I doubt you're keeping track, but Donovan and I hope to add a few more cites to your tally with the papers that we have under review. 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 our seminar series and data products we've created for dementia research.
Music and engineering for this podcast was provided by Dan Langa. More information available at www.danlanga.com. Minding Memory is part of the Michigan Medicine Podcast Network. Find more shows at uofmhealth.org/podcast. 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 represents the views of the NIH or the University of Michigan. Thanks for joining us. We'll be back soon.
Listen to more Minding Memory podcasts - a part of the Michigan Medicine Podcast Network.
