The Stats Behind Aging | Stats + Stories Episode 94 / by Stats Stories

Dec 2013 headshot.png

Karen Bandeen-Roche, PhD is a statistician / gerontologist who currently chairs the Department of Biostatistics at the Johns Hopkins Bloomberg School of Public Health. Her statistical specialty areas of research are in latent variable and multivariate outcome modeling. Her gerontologic research aims to better understand the causes and course of physical disability, cognitive decline, and frailty in older adults, so that their adverse implications can be delayed or avoided. She is an ASA Fellow and a Marvin Zelen Leadership in Statistical Science Award winner, and she has contributed extensive service to promote the statistical profession through leadership in scientific review panels and our professional societies.

+ Full Transcript

Rosemary Pennington: The term Gerontology was coined in 1903 to describe the study of aging. It’s a multidisciplinary field, including scholars interested in the cultural or social aspects of aging, as well as those whose interest lies more in the realm of biology or psychology. One area of interest for scholars is the issue of frailty in older adults. The study of aging and frailty is the focus of this episode of Stats and Stories, where we explore the statistics behind the stories, and the stories behind the statistics. I’m Rosemary Pennington. Stats and Stories is a production of Miami University’s Departments of Statistics and Media, Journalism, and Film, as well as the American Statistical Association. Joining me in the studio are our regular panelists, John Bailer, Chair of Miami Statistics Department, and Richard Campbell, Former and Founding Chair of Media, Journalism, and Film. Our guest today is Karen Bandeen-Rouche. Bandeen-Rouche is the Hurley Dorrier Professor, and Chair of the Department of Biostatistics at John Hopkins University. One of her areas of expertise is the study of frailty. Karen, thank you so much for being here today.

Karen Bandeen-Rouche: It’s my pleasure. Thank you.

Pennington: Can you just explain what frailty means in the context of gerontology?

Bandeen-Rouche: Well, at a geriatric level, frailty has been recognized, probably, since there have been older people. There’s a quote that Shakespeare had in, “As You Like It”, referring to the older man whose hose had shrunk and no longer fit his shrunk shrank, but it’s only about – in the last 30 years that the study of frailty has become rigorous and standardized, and from that point of view, I would define frailty as a biological syndrome of vulnerability to adverse events, following on stressors that arises due to dysregulation of multiple physiological systems that govern our overall homeostasis and stress response.

John Bailer: Wow [LAUGHTER]. That’s a great definition. There’s so many components. So, one of the things that’s a natural question when you think about this is to try to parse this into pieces. So, let’s take a stroll through that definition again, Karen. So, you talked about some of the – idea of vulnerability, so I think of that as, kind of, the likelihood, the probability that something bad can happen, some kind of chronic disease, some kind of injury, mortality. I sort of hang with you there, but can you talk about some of the other dimensions that you mention in that definition?

Bandeen-Rouche: Well, absolutely, I mean, and so the other dimensions, I think, have to do with trying to get at what the underpinnings or the etiology of frailty is, and there’s a school of thought that would hold that it’s not just the same as old age, it’s not just the same as having a lot of diseases, but it’s truly reflective of a point of crisis in the interactions between the systems in our physiology or maybe erosion within those systems themselves, and the interesting thing about this, from a statistical or mathematical point of view is that, then we can almost think of the body as a machine and frailty then, becomes a state of the whole system of components and how they interact that we might be able to characterize mathematically.

Campbell: One of the things that you try to do in your work is to sort of get people to separate physical causes from cognitive causes. So -- and I’m thinking here of practical example. My own mother, who’s 92, who’s frail but very sharp mentally, but I also have someone in our family who has a Parkinson’s Plus degenerative brain disorder, that causes her physical problems. So, she’s only in her mid-60’s, but she’s also – she’s frail because her gait is slow, she’s exhausted. Some of these things that are physical manifestations are happening to her, but there, the cognitive disorder is causing her frailty. So, I mean, I’m trying to understand this from a real practical point of view, and I know you talk in your discussions about the implications of this, and you talk about prevention, you talk about intervention. So, I think there was a question in there. [LAUGHTER] Could you talk a little bit more? I’m real interested in this relationship between cognitive and physical.

Bandeen-Rouche: Yes, and you actually remind me of a conversation I had with my older brother recently, in which he basically asked me, why are you so pedantic about frailty? [LAUGHTER] He had seen some other video of me online where I was talking about the distinction between physical aspects of frailty and cognitive aspects of frailty. And it is true, that the group that I’m a part of has really focused on physical frailty. And so, everything that I just said, was closely tied up around that. But I would totally agree with you, that there are so many other things that can contribute to an overall loss of ability to cope, or compensate, or to be resilient for older adults in their daily life, and certainly, cognition is one of them. And so, cognition could, I believe, both cause a separate dimension of vulnerability, inability to manage stressors. But you’re right, it could also drive what some of us, technically call, secondary frailty, which is the idea that absolutely cognitive decline, either directly or maybe indirectly, by its effects on one’s physical health, might then drive changes that manifest in all of the same symptoms that you would expect from physical frailty. I think that there are colleagues of mine are interested in both directions of that relationship actually. The idea that, certainly, cognitive decline may cause either frailty itself or symptoms of frailty, but also the reverse, that is, one becomes frail physically, one may then not be able to engage in physical activity, or healthy sleep, or other activities that maintain our cognition, and so it could really be a very global, vicious cycle that results.

Bailer: You know, you talked about this – we’re talking about this more generally, I’d like to dive into some of the specifics of some of the dimensions or the things that you actually measure to operationalize frailty. I looked at some of the criteria which included, exhaustion, low physical activity, weakness, slowness, and shrinking, and can you talk about how those – why those were determined. Why were those five things selected as part of this criteria, and how is that then ultimately mapped to the idea of being robust, versus pre-frail, versus frail?

Bandeen-Rouche: So, yeah, that’s an interesting question, and I – those five criteria were chosen, I believe, both from a geriatrician’s point of view, meaning decades and generations of observation, of the sorts, of properties that frail, older individuals manifest as they come to a point of being ready to fall over a cliff, as well as from the etiology that I spoke of, the multi-system dysregulation. So, if I begin with the latter, the system specifically that are hypothesized to be involved, are those that govern energy production, and the use of energy, and then stress response. So, for example, inflammation we know is a process that is healthy, under the right circumstances. When we exercise, there’s a little tiny bits of muscle damage, and so the inflammatory system kicks in and executes repair, but in older adults, that process can be turned on all the time and actually results in muscle wasting, and in other processes being messed up, because inflammatory markers bind to molecules that are supposed to be bound by other things. And so, the energetics part of it is certainly reflected in exhaustion and low physical activity. The inflammatory hypothesis is reflected in weakness, having to do with muscle wasting, also weight-loss, and then slow gait speed is thought of as being an integrative marker that includes some of the things I just described, but also the idea of slowing in motor pathways. And so, this goes to the previous question, that actually there is a brain component there which governs motor function. And so, that’s the etiological side. And on the geriatric side, I think that all five of those things would be symptoms that geriatricians also would think of as they manage their patients and worry about them becoming frail.

Campbell: Speaking of geriatricians, is that how you say that word?

Bandeen-Rouche: Yes, yes.

Campbell: So, how have – you talk about your research being used, possibly, for intervention and reversal of frailty, how do doctors respond to your work, because they’re the ones that are going to have to manage this. And is it a, kind of a slog explaining it to doctors?

Bandeen-Rouche: I think there’s a tremendous interest. Certainly, the literature on frailty has skyrocketed in the last ten years, and clinicians in specialties other than geriatrics, have really latched on to the idea, particularly in fields like cardiology, surgery, nephrology. There’s an extremely act of work going on, I think, because of the promise of potentially being able to identify older adults who really may be vulnerable to interventions that are needed in those fields. Or maybe the converse, if an older adult is robust, there is an individual who shouldn’t be excluded from a potentially beneficial procedure just because they’re old. And so, there’s a tremendous amount of interest and such. Physicians are starting to measure frailty, but I’ll be very honest that I think the field is still very early on in being able to make informed recommendations about exactly what one should do when a person is recognized as being frail. And that’s what’s so exciting about the present moment, both the development of care management strategies and plans and executing randomized studies to evaluate their efficacy when frailty is used as a mean to target treatments to people. And then on the flipside, better understanding the etiologies, so that one might intervene on that directly and prevent or delay frailty in the first place.

Pennington: You are listening to Stats and Stories, and today we’re talking to, John Hopkins University’s, Karen Bandeen-Rouche. You were talking about, sort of, where the field is now, how has it changed since you entered?

Bandeen-Rouche: I think the main way that it’s changed is that there has been this skyrocketing interest in the field, and then along with it, something that some consider to be beneficial, and I have more mixed feelings about, which is an explosion of frailty instruments and methods that have proliferated. And so, the benefit, of course, is that different ways of measuring frailty become available to clinicians, and a given clinician may be able to identify a method that works well in their particular setting. The drawback, of course, is if measurement is not standardized, then it becomes difficult to rigorously evaluate either etiology or care management strategies that then generalize across methods or even to be confident that strategies are identifying the same sets of people. So, it’s a moment, both, of great excitement, but of some amount of debate and controversy in the field as well.

Bailer: You know, as I look at how you describe this, and some of the ways this is measured, it appears that there’s this transition model that you go from robust to pre-frail, to frail, in transition, in these conditions, and is it possible to move back along this pathway, to move from frail to pre-frail, or from pre-frail to robust? Are there interventions, or have you looked at studies that have tried to do this, and have they been successful?

Bandeen-Rouche: There are a limited number of studies, and most of them, interestingly, from a statistician’s point of view, have intervened directly on the criteria. And so, for example, implemented a physical activity regiment to address slowness and lack of physical activity or implemented stress training to operate on weakness. And so, there are some of those interventions which have resulted in a technical reversal, say, from frail to pre-frail or pre-frail to robust. The concern that I have about many of them is that much of the change has been observed in the physical activity criterion, subsequent to a physical activity intervention. And so, I – my own view is that there have not been many, if any, intervention studies that have demonstrated a reversal of what I would consider to be the underlying drivers of the symptoms by which we recognize frailty, and that is a very active element of, I think, the immediate front here for frailty research.

Campbell: So, frailty research, I want to know how you got interested in this. You’ve got a BS in Mathematics, a PhD in Industrial Engineering from Cornell, so what got you to this point?

Bandeen-Rouche: At one level, it was just luck. [LAUGHTER] I came out of graduate school and my interests all along have been in complex measurement, that is the idea of, how do we measure, or perhaps evaluate proposed measurements of phenomena that are well-describable, that have a cogent theory that characterize what one, ideally, would like to measure but is somehow infeasible to measure and can only be measured via indirect surrogates or indicators. And so, when I came out of graduate school and I came here to John Hopkins and as in biostatistics departments, one meets people and looks for collaborators in which there is a good personality and match of interest to then work together, and I just stumbled upon aging as one of these areas, but before long, it really hooked me, both because it was a great match to this measurement issue, but also because the problems are so compelling and the potential to positively impact society with the skyrocketing population of older adults, is tremendous. And so, it was probably ten years thereafter that the notion of frailty, as a more rigorously-theorized and standardized construct, was beginning to emerge in the gerontological literature, and both, from the point of my statistical interests, but also for the promise of frailty as a point of the sphere by which to identify vulnerable individuals, better manage their care thereafter, or to prevent a cause of subsequent decline altogether, was extremely appealing. And so, I would say that, that’s basically how I got into it.

Campbell: Very good.

Bailer: I was intrigued at some of the work that you had done, looking at some of these predictors of frailty and just seeing things like, high income being associated with high robustness, or race ethnicity differences and robustness frailty, or particularly, the regional differences in the United States in terms of the percentage of frail individuals. So, can you talk a little bit about, kind of, what you’ve learned or what you think about some of the reasons why there’s this difference across the US, in terms of frailty.

Bandeen-Rouche: You know, we were fascinated by that as well. I believe that the extent, particularly of racial and regional disparities, really surprised us, even though we, in some sense, expected them. And so, I guess our leading hypothesis go, both, to causes and to measurement. And so, on the side of causes, I, both under – individuals who may be disadvantaged or just cultural patterns in some regions of the country, may be tied to a less healthy lifestyle than for others, and that could well drive changes in health that, ultimately, are just manifested in greater rates of frailty. And I believe we all expect that that would be a part of it, but we’re, right now, in the middle of some subsequent work that is evaluating measurement differences as well. And so, we have identified, for example, some very large hints that – walking speed, for example, which we think of as an objective measure, may actually be tied to our personal characteristics more than we think. So, for example, cultural differences in usual gait speed. Exhaustion is a self-reported item, and we’ve seen some real hints that individuals of the same health status but with different personal characteristics may be either, more or less, willing to report exhaustion, or think of exhaustion as a somewhat different thing and so may report it differently. And so, these are definitely findings that we also have found very interesting or trying to follow up on right now.

Campbell: So, one of the things that you mentioned, just in terms of talking about measurement is the imperfect measures, and it just reminded me, what are your biggest challenges as a researcher in this area? What kinds of things would you like to know that you feel like you can’t, kind of, get a handle on or anything else that might be challenging to you and doing this kind of work?

Bandeen-Rouche: Yeah, on the measurement side, part of it is what I just talked about, you know, these emergent hints that the measures that we have, may perform substantially differently, and people with different personal characteristics, to better understand why that is and to then move ahead and potentially identify a next generation of measures that performs better. I think that, that is the thing on the measurement side that is the most compelling and needful for my point of view. And then of course, on the etiology side, just learning what are the causal drivers of changes that manifest in frailty in some older adults and not others, I think, is a wonderful biological puzzle, and it’s also a mathematical puzzle, and the intersection of those two things makes the problem absolutely fascinating.

Bailer: You know, given what you just observed about the idea of cultural differences and some of the ways these measures are generated, now I’m curious about what would happen internationally. Do you have measures – does this measure of frailty, that you have, would it apply if you were looking at other communities around the world?

Bandeen-Rouche: No, so, this is a question that has generated a huge amount of interest in the gerontological community as well. Certainly, these measures are applied around the world, but concerns have, indeed, been raised about the homogeneity of their application, and the observation that, people in different countries may tend to have, for example, different walking speeds, or may tend to think of a concept like exhaustion differently. And so, there are debates about, well, should you create different cut-points to define what would meet a criterion of frailty on a given indicator in different countries? And then that, of course, raises questions of a statisticians, should we go along with a clinical desire to create cut-points in the first place, as opposed to, taking the measures and whatever richness they come to us, and analyzing what might be a more optimal way to characterize subgroups of robust, pre-frail, or frail. And so, those are all great questions for the field too.

Bailer: I’m curious, geronmetrics, is a phrase that I’ve seen you use. Did you start that? Are you the initiator of that term?

Bandeen-Rouche: That’s a great question, and so, the truth is that I’m not sure. [LAUGHTER] I thought I had.

Bailer: Well, we’ll credit you with it.

Bandeen-Rouche: Well, I don’t know, but the other people who should be considered for that designation are the biostatisticians and the Claude Pepper Older Americans Independence Center at Yale. Those are the other group of people that I’ve seen use it a lot, and the truth is, I’m not positive that I used it first, but I don’t remember hearing it before I used it.

Bailer: So, let me follow up just real quick, what would you advise someone who wants to become a geronmetrician, someone who’s interested in being involved in the kind of work that you do. What type of preparation would you suggest for getting into this world?

Bandeen-Rouche: Well, I think the thing that has helped me the most is just to dive in deep into the science itself. It’s a field where researchers are incredibly interdisciplinary, and welcoming, and in particular, welcome statisticians to sit at the table as a full scientist along with every other scientist at the table. And so, if one digs in and learns the science, and cares about the science, there’s an unlimited potential to lead in the field, and then the field itself presents so many diverse challenges that I believe that, virtually, any interest in statistics has a vital place to make a difference in gerontology.

Campbell: I also just want to compliment you for starting your podcast with us by quoting Shakespeare. [LAUGHTER] I always appreciate when statisticians and mathematicians quote Shakespeare. Well done.

Bandeen-Rouche: I did my best.

Bailer: You just won Richard over right there, Karen. [LAUGHTER]

Pennington: Karen, that is all the time we have for this episode of Stats and Stories. Thank you so much for being here.

Bandeen-Rouche: Thank you. I really enjoyed it.

Pennington: Stats and Stories is a partnership between Miami University’s departments of Statistics and Media, Journalism, and Film, and the American Statistical Association.

You can follow us on Twitter, Apple Podcast, or now Spotify. If you’d like to share your thoughts on the program, send your email to, or check us out at,, and be sure to listen for future editions of Stats and Stories, where we discuss the statistics behind the stories and the stories behind the statistics.