Elder Abuse: How Much Occurs and How Do We Measure It?

Introduction: Carrie Mulford, Program Manager for the Elder Abuse Portfolio

Carrie Mulford: So I'm going to be very, very brief. Welcome to this session; it's "Elder Abuse — How Much Occurs and How Do We Measure It?" — just to let you know if you're in the right place. I am Carrie Mulford, I am the program manager for our elder abuse portfolio at the moment, and it's been a very growing and exciting portfolio of research that we've been developing over the past four or five years. And one of the areas that we're really focused on is this issue measurement - and the issue of measurement is important because if we don't know how to measure elder mistreatment then we - there are two major things: one is it's very difficult to make a case to policy makers that this is an important issue that you need to pay attention to; and the other is if we don't know how to measure it then it's very difficult to assess cases as they are brought to the attention of APS, law enforcement, etc.

So with that, I'm going to just introduce the speakers and turn it over to them. The first person to speak is going to be Ron Acierno, and he is the director of the post-traumatic stress disorder clinical team at the Ralph H. Johnson VA Medical Center and an associate professor of in the department of psychiatry at the Medical University of South Carolina. He's been involved in elder abuse research both through the National Institute of Justice and the National Institute on Aging. And he's going to talk to you about a prevalence and incidence study that he did for NIJ and he's currently the principal investigator on that study.

Our second speaker is going to be Ken Conrad, who is a professor of public health at the University of Illinois at Chicago where he focuses on measurement, theory and practice. His interests are in long-term care, homelessness, mental health issues, substance abuse and, in general, measurement and evaluation issues. He's the principal investigator on an NIJ study that's just wrapping up — on measuring psychological and financial abuse using what probably to a lot of you will be a new measurement technique, and he'll talk to you about that.

[Audio of Ken Conrad's presentation is not available.]

And then we are very happy to have Georgia Anetzberger here from — she's an assistant professor in the healthcare administration program at Cleveland State University, and I always consider her to be extremely insightful and practical, so I love to have her as either a discussant or participant in pretty much anything that we do related to elder abuse. She's also been involved in the field of aging and human services for more than 30 years, so she has a lot of experience to draw on in tying all these things together. So, with that, I am going to turn things over — I will be keeping time and doing nothing else.

Presenter: Georgia Anetzberger, Ph.D., Assistant Professor, Health Care Administration Program, Cleveland State University, Ohio

Georgia Anetzberger: Good morning. I was asked to be the discussant on both of these studies, but I want to, in a sense, start with a larger context. There's been a long-term interest in counting how much elder abuse goes on and figuring out how to do that. And I would argue in fact that there's been a half century of interest in this regard. It goes back — in a sense — to the earliest national discussions on elder abuse, and those happened in 1960 at the Arden House Conference on Aging, where there was a concern about older adults that might potentially either neglect or be subjected to exploitation– and a desire to develop protective care to deal with them. But there was recognition at that time that nobody knew how many such individuals there were out there.

A quarter century later at the first elder abuse research forum, which was held at the University of New Hampshire in 1986, people lamented that they still didn't know — not just that they didn't know how much elder abuse went on, but that they didn't know a lot of other things, that they didn't know how to get there in terms of methodology. And finally, more recently, in 2003 at the National Research Council's panel to review risk and prevalence on elder abuse — this is the report that came out, and this is what has kind of propelled studies like this, as well as other studies — there was acknowledgement that there was insufficient information about the occurrence of elder abuse across all kinds of settings that older people might be found in.

Now that doesn't mean that there hasn't been prevalence studies that have been accomplished. And the earliest, in fact, occurred in 1979 in Maryland, and it was done by Marilyn Block and Jan Senate. They sent out letters to a random sample of older adults in that state — sent out over 400 letters — and they got back 73 responses and came up with a prevalence rate — looking at four different kinds of elder abuse — of 4.1 percent. The classic example of prevalence studies, and the best regarded, has been that done by Carl Pelmer and David Finkelhore, published in 1988, looking at the metropolitan Boston area — a random sample study again. 2,020 older people were investigated, looking at three different forms of elder abuse, and a prevalence rate of 3.2 percent.

And finally, this one has already been referenced much more recently — published in 2008 — Edward Lowman and his associates had a nationally representative sample, looking at just three elder abuse forms, with the highest rate going towards emotional abuse at 9 percent.

Throughout this entire period, there has been recognition about research limitations in coming to grips with frequency of the problem. There's been concerns about definitions, lack of a nationally accepted standard in terms of definitions — all kinds of concerns about what are the recognized forms of elder abuse. There's been an acknowledgement about flaws in research design as well as ethical questions on even how to even conduct this type of research, especially with cognitively impaired older people. And finally there's been acknowledgement of all kinds of sampling inadequacies from very small samples to very biased samples.

Now this isn't to discount the fact that it's really important to figure out prevalence. And it's important for at least three reasons — one of which was acknowledged by Carrie just a moment ago; and that is, somehow you cannot interest legislators in terms of public policy unless somehow you show that it's a very big problem; they're not going to be interested in public policy around impacting this problem unless it's sizeable in terms of impacting older people. It's also unlikely you're going to get the deployment of resources — either from government agencies or from other organizations — unless a number of people are impacted. And I'll give you an example in terms of my own state of Ohio. On two different occasions, historically, we have lost all of our funding for adult protective services, and all of those dollars have gone to child protective services because we couldn't sufficiently show a large number of older people being affected by the problem — and we're likely to encounter this again this year.

And finally, you only grab media attention and therefore increase public awareness about a problem if you can show it's either so rare, so unusual or so horrific — which of course many instances of elder abuse are — or if you can show it's a really big thing and therefore deserving of media attention.

Now even if we know the prevalence of elder abuse, it's not going to answer all the questions that are out there. It's not going to tell us the underlying ideology of the problem. And if we don't know why elder abuse occurs, we aren't going to know how to either prevent the occurrence of this problem or its reoccurrence. Knowing the prevalence of elder abuse is not going to tell us anything about the consequences of abuse affliction. It's not going to tell us the suffering, the injury, the disability encountered by elder abuse victims. And it's not going to tell us the cost to society of having this problem in our mist — the cost of healthcare and other sorts of things. And finally, knowing the prevalence of elder abuse is not going to be helpful in telling us effective detection and interventions. It's not going to tell us what works and what doesn't work in trying to impact the problem.

That said, I want to talk just a little bit about each of these two studies, beginning with Ken's. I think Ken's is a very important study in terms of getting at how do you define certain elder abuse forms, and obviously in this case it's financial exploitation — and although he didn't discuss it, psychological abuse. Because what's so important about the work that he has done is that he really begins to fare it out — the dimensions, in detail — the dimensions of this problem. And he does it using rather sophisticated research designs and techniques for analysis. One of the things that has been desired in the field of elder abuse for a couple decades now is to get individuals who are researchers who are already established and perhaps concerned with other areas of research — get them into the field of elder abuse so that they can make a contribution to this field, using techniques, using research methodologies that have been found useful in other subject areas; and clearly this is what Ken has been able to do in his particular work with respect to financial exploitation and psychological abuse.

Now that doesn't mean that there aren't some issues with respect to this particular study, and I've just identified a couple of them, and a lot of them to me have to do with the sampling issues with respect to the clients. For example, probably not the best interviewers, with respect to the clients in a situation like this, are the adult protective service investigators; there's a lot of potential bias in that particular regard. Also, in terms of the client population of 227 individuals, 61 percent were African Americans — in the general population of older people in this country, it's less than 9 percent. That's a very heavy bias with respect to one racial grouping. On the other hand, that particular racial grouping — African Americans — have been shown in a number of other studies to be particularly attuned to the problem of financial exploitation, to be particularly concerned about it, and having more experience with it than other racial groups.

And finally, in terms of the concept of substantiated financial exploitation and psychological abuse, staff marked it as substantiated even when there was only some indication of it which would have to do with signs or symptoms or that kind of thing. And as we all know, that might necessarily deal with other things and not elder abuse.

There were a number of conclusions that were interesting, and obviously I can only highlight on a couple of them, so let me do that. Among the different findings that come out of Ken's studies, is the fact that the structure of the client data represents a much simpler picture of financial exploitation than did the data coming out of experts — and I would suggest that's to be expected. The experts always look at the detail — at the richness of something — because that's their work, it' their area of concern. And the classic example that I can think of in that regard — my first master's degree happens to be in anthropology, and so the classic example in using those anthropological perspectives is: I come from the Snowbelt of Ohio, so I'm very familiar with snow. On average, we get about 105 inches a year, but I only know about three different words for snow — only about three different variations, so to speak; where Eskimos supposedly know more than 40 because they're the experts; it's a matter of their survival, their life.

Another conclusion that comes out of this study is that the clients saw coercion as more severe, and abusive trust as less severe than the experts; and I would say that's to be expected as well. Clients have a complicated and long-term relationship with their perpetrators, and therefore the concept of abusive trust would be less compelling to them then coercion. And let me give you a concrete example. If I have my adult son go out to the store for me and I give him a $20 bill to buy me a loaf of bread, and he comes back with the loaf of bread but doesn't come back with the change because he used all that change to buy lottery tickets for himself, I'm going to be sad and I'm going to be disappointed, but we have the kind of relationship, to be truthful, that I realize that it's more than just that particular incident. On the other hand, if he coerced me out of the $20, it would be such a violation of my perceived rights as an individual — it would be such a violation of my self determination — that I would be very much concerned.

Let me go to Ron's study. Ron's is really important for a couple reasons. First of all, it is the largest nationally representative study that has taken place with respect to elder abuse. The one by Lowman and his associates had as a sample 3,005 individuals, and here we have Ron, who has exceeded his own expectations of targeting 4,000, by getting 5,777 in his sample. In addition, he looks at more elder abuse forms than any of the other prevalence studies have done. The third reasons that this is so important is because Ron has a history of looking — using this kind of methodology — at younger abuse populations or younger populations overall. And what's nice about having now done it with an older population is you can begin to do comparisons across age groups, and using Martha Stewart's, it's a very good thing.

That doesn't mean that there might be some issues, and let me give you one example of that. I had some concern about how he defined certain sorts of things that seem to be kind of representative of the population as a whole. He terms low household income as under $35,000 a year, and yet that represents nearly the majority of the individuals in this representative sample. He terms very low levels of social support, and yet again this is nearly the majority of the individuals in his sample.

There are a number of conclusions of interest. First of all, as you saw on the slide when he was coming to an end of his presentation, approximately one in ten individuals who are community-dwelling older adults had experienced some sort of elder abuse in the past year, and of course it increases to one out of seven if you include financial exploitation. Generally speaking, before this study, people would use a kind of range of elder abuse occurrence between 2 and 10 percent. So what we're finding out with Ron's study is it's beyond expectations in the past — and, in fact, it's really at the upper limits of those expectations. And it almost harkens back to the earliest research on protective services, going back 45, 40 years, where they came up with the number of people that might be needing a protective intervention as somewhere between 15 and 20 percent, so it's much more of that caliber.

The second thing he found was low levels of social support correlated across almost all elder abuse forms. I would suggest that low levels of social support, lack of social support, social isolation, have been recognized as salient risk factors for abuse occurrence for over 20 years now and have been found to be true in a number of studies. What's particularly interesting in this regard, however, is that we are seeing lower levels of individual social support among Americans overall, and this has been going on over the last 20 years. A study done in 1985 found that the typical American had three close friends; more recent studies are finding that it's two close friends. We have only to look to the future in terms of our aging society to wonder the implications of that for the future of elder abuse.

Another finding on Ron's part was that emotional, physical and sexual abuse have a kind of perpetrator profile that includes high unemployment, increased substance abuse, increased likelihood of mental health problems and social isolation. There too there's have been a number of studies going back 20 years that have found this to be almost the classic profile of the elder abuse perpetrator — and my own research included in that regard. And I would suggest that Ron's absolutely right with this kind of a profile, there are things that can be done in terms of rehabilitative, social and other services, to deal with this, to prevent occurrence, to prevent reoccurrence. And yet I was surprised then, when Ron came up with the following concluding statement, and he did — I saw it in your slide, but you didn't in what you had sent me — and what it said was, and I'll quote, “Unfortunately most of the mistreatment would not qualify for criminal justice prosecution under most current statutes. This factuality is disappointing and speaks to a social acceptance of abusive behavior.”

Well, I would argue that elder abuse has been conceived in such broad terms that there's not going to be this single intervention — criminal prosecution or otherwise — that's going to be able to deal with all forms of elder abuse. Rather, criminal prosecution will be one of many that needs to deal with it. And elements like certain forms of neglect, certain forms of violations of rights can be dealt with in other channels, through other means. Thank you.

[Applause]

Presenter: Ron Acierno, Ph.D., Associate Professor, Department of Psychiatry, Medical University of South Carolina, Charleston

Ron Acierno:Hello. This talk is going to be sort of an epidemiological talk — a sort of bird's-eye view of a study we just finished — probably a little bit cart-before-the-horse because there is still a lot of definitional debate and things like that. But we had several pilot studies that said we could probably do it so we wanted to go ahead and do it. This is the research team at the Medical University at South Carolina primarily, and SRBI was our survey research firm. Just a couple of goals: just want to go quickly over methodology, some decision points we made, then talk about the study and its findings.

Two conceptual issues that we had to consider before we started were sort of how we're going to address elder abuse. The first methodological turning point in the road is: Do you just look at abuse by family or caretakers or do you also include assault, typically — stereotypically defined as something by strangers. And the other one: is the model you're going to use based on cognitive status? Because that should also determine your methodology. So most existing studies only look at one or the other — abuse or assault — and the problem is that they share common sets of risk factors, and they also probably diverge in terms of preventive intervention so it's probably a good idea to study both if you can. Our methodology that we piloted with younger adults and adolescents and into older adult samples sort of did that — and then conceptualizing elder mistreatment in terms of cognitive status. If you use a child abuse model, it might be problematic, since most of the elder abuse we know about or talk about for noncognitively impaired adults is sort of spousal in nature or adult-on-adult. The other problem is that the methodology is going to differ. Child abuse methodology a lot of times looks at sentinel reports as opposed to direct interviews of the child, whereas adult interpersonal violence assessment is a direct interview of the index person.

We sort of thought it looked more like domestic violence, younger adult interpersonal violence and child abuse because of the cognitive status of the older adult. They're independent older adults; typically the abuse is one independent adult against another independent adult, and not a significant power differential like in child abuse — and again we're talking about noncognitively impaired older adults. And then you have cognitive development — there are two mature adults. So we realize that if you are going to study cognitively impaired older adults, you probably do want to use more of the child abuse model, but if you're studying interpersonal violence in cognitively unimpaired older adults, then the way we think to go is sort of the model we used on younger adults, which is direct respondent interview.

OK, let me just catch up to myself. Just wanted to give a quick overview on methodology — that's not the core part of the talk but to let you know why we're focusing on cognitive unimpaired, community-residing older adults. Basically we used random digit dial methodology to study almost 6,000 older adults. We used computer assisted scripts, where the same questions are asked of every participant. If they say “yes” to an item, then it will go into further depth on that item. If they say “no,” it skips out so the respondent burden — the time it takes to interview people — is much less.

We had a very good cooperation rate, as is typical in a lot of older adult studies, of almost 70 percent of eligible people, and the participants were interviewed in either English or Spanish at their preference.

The mistreatment variables we looked at were emotional, physical, sexual, neglectful and financial mistreatment, and for various reasons we had to keep neglect or potential neglect toward past-year events and financial exploitation by family toward past-year events — the others we got since age 60, lifetime and past-year. And then we looked at some risk and protective variables — demographic factors such as age, gender, income; self-rated health status, to look at sort of the idea that people with poor health are increased burdens on, for example, the people in their environment might elicit — through no fault of their own — more mistreatment behavior; perceived social support — which in some of our research with older adults and disaster situations seems to be a very important factor; use of social services; and level of functional impairment.

As I said, there were almost 6,000 people aged 60 and over; we also interviewed almost 1,000 proxies. Now these are people who either helped care for or who lived with an older adult. The reason we did that is we were trying to test a method of determining violence prevalence by proxy and cross-checking that with the respondent interview just to see if it was a valid index. Previous studies have actually had more luck asking the people in their environment — of the older adult — if they had been abusive toward them than they had with asking the older adults themselves. And in this study, that didn't really work out, but that's an entirely separate talk; it's something we'll have to revisit. Some more women than men — as you would expect at this age because men don't seem to last as long; and almost two-thirds married. We mapped pretty closely to the demographics of the population, with only minimal correction.

And then some of the risk factors we used — just to give you an idea of the context: Older adults are an economically at-risk population, being on a controlled income with increased costs of living. Almost half of our sample were in a household that had less than $35,000 per year — and that's the entire household. We did expect a lot to be retired. Only a fifth would consider themselves in poor health for their age — that's sort of an age-based “do you consider yourself in poor health for your age?” Two-thirds had a prior traumatic event; this would be a criteria for a post-traumatic stress disorder stressor — something really bad. About half thought they had lower social support than they needed. And they were using some community or social services so there was some need and some access that was happening; that was a little reassuring. And some needed assistance with daily living — about a third.

OK, going right into the variables: This is the area of consternation, is how we define these things. So what we did is we tried to — after a couple of our pilot studies — bridge the literatures of younger adult and adolescent interpersonal mistreatment and extend it to older adults to have a starting point rather than get in-the-field consensus, which I think is going on, on the CDC side. So basically, the way you do this that we found with the younger adult literature is you have to have a contextually orienting preface statement, which you follow up by yes or no questions, as opposed to open ended questions — sort of the opposite of a clinical interview. The contextually orienting preface statement gets people away from the stereotypical response set, which is typically recent events which were perpetrated by strangers that you told the police about. That's the default response set. So we want to sort of undo that set by saying we want to know about recent and distant past events, whether or not you reported them and whether or not you knew the perpetrator. In younger adult populations, you get a 1,100 percent increased hit rate when you do that.

So here's the questions we asked specifically: Has anyone ever verbally attacked, scolded, yelled at you? Threatened you? Intimidated you? Made you feel so humiliated or embarrassed? Calling you names, making you feel worthless? Repeatedly coercive requests to do something over and over again? Or completely refusing and isolating from a person? These would be considered emotionally abusive events — perhaps not in violation of a law in most statutes, but nonetheless not very nice.

Here's what we had in terms of our emotional mistreatment prevalence. The red is past-year, so around 5 percent. Lifetime, it was up to 20 percent, and you would expect somebody to be mean to you over your life. And then the yellow is — or the gold is — since age 60, so that's abuse happening during older adulthood. And these are the subtypes — verbal humiliation and harassment are being ignored. The gender splits were pretty even; there's no significant difference.

This is where it starts to get interesting: Who's doing it? Well the stereotypical abusive stranger doesn't do it that often, and that's consistent with the young adult literature. It's typically people in your environment that you sort of expect would be nice to you — partners, spouses and acquaintances, children share the load there. Just some characteristics — some of these numbers may be hard to see: Most is not reported to police — less than 10 percent. Most is by family member, but the thing that is of interest to us is the perpetrator characteristics. Half are unemployed; a fifth had substance abuse, mental health history and social isolation.

These are areas of potential intervention, and you're going to see this consistently throughout the perpetrator classifications for each mistreatment type. And so we think that addressing some of these things may indirectly address elder abuse. Now this slide is just very busy — these are all our individual risk factor chi scores. I put it here — the significant risk factors are here, and the ones that are in red sustained their significance in logistic regressions — meaning they independently contributed to risk. In this case, in contrast to prior research about the older old — the eldest old being at the highest risk and consistent with the recent study by [Edward] Lowman, lower age was a big risk factor for us. So our younger older adults were more likely to be yelled at or emotionally abused. Being employed — now that was unexpected. Basically, some of this stuff is happening in the work place — we thought that would be a protective factor, but it wasn't. Health dropped out, but experiencing prior traumatic events stayed there. So basically these people come from a potentially traumatic context — again though, half the population has that event so we can't put too much weight on there. Low social support — remember that one. Low social support in my psychiatric epidemiology research is a predictor of depression, PTSD, all sorts of bad things. Well it seems to also be sort of a risk factor for emotional abuse — something that we may want to target. And then needing help with daily living — so there's your risk factors there.

Moving on to physical mistreatment — here's how we defined it. We sort of set it as a behaviorally specific event. Instead of saying, “has somebody physically mistreated you,” we actually try and use behavioral terms to lay it out, and then we asked these specific questions. So now again if you ask yes or no questions and you're on the phone — it sort of increases the safety of the individual, because even if the perpetrator is in their environment, all they hear is a yes or no question. It also makes it easier for the person to answer without having to go into what happened to them — it's less embarrassing, a little bit more anonymous.

So these are the three types of physical abuse we asked about — now these should actually be illegal on most statutes. And here's what we found; this is much higher than Lowman's study, and I'll put their bar graphs up next to mine in a second. They just completed theirs about a year ago — two years ago I guess. Almost 2 percent overall, and in an even spread hitting, restraining and injuring, or assaulting you to be injured — and this is past-year. The gender split was not significantly different. The perpetrators — again, strangers are not doing this to people; it's the person in their environment. This is mapping on very nicely to the younger adult interpersonal violence literature but even more so in terms of familial and acquaintance aggression. Most of it is domestic violence it looks like so the domestic violence model is going to be relevant to the cognitively intact older adults. Very little is reported to police — more than younger adults but still not as much as one would hope for physical assault. Now sexual assault, nobody reports anything to the police, but physical assault you would hope a little bit more. In younger adults, it's about 50 percent; in our older adult population it was only 30 percent — again that's probably because it's the family members doing it. This family member may drive them to work; this family member may be the person keeping them out of a residential care facility.

Again, areas of intervention ... Substance abuse of the perpetrators — that's half of the known perpetrators were known to have substance abuse by the victim. Unemployment — we should expect that to get worse. Unfortunately, that's a risk factor for physical assault on this population so that might have a connection to what we're going to see in a few years. And then social isolation on the part of the perpetrator and mental health history. So again, areas of intervention, that's sort of the positive aspect here.

Again, everything here was a risk factor in univariate analyses, and then when we considered everything together, the things in red floated to the top. Lower age again, and again low social support. If a person perceives that they didn't have someone to talk to, that they couldn't go to somebody when they needed help, somehow this also contributed to increased risk of physical assault. Maybe connected by proxy to social isolation, which we know is a risk factor from prior studies.

Switching to sexual mistreatment — this one we had tremendous difficulty in our Spanish translations because whereas the pilot testing in English, very few people were offended, very many people were offended in the Spanish version, so we had to work on this a little bit. Again we set the context very clearly so that it could be somebody that you know or didn't know, whether or not you reported it to the police, whether it happened recently or in the distant past. And these are the areas again that should be illegal in most states. Forced sex, forced molestation, forced undressing. These rates are encouragingly low — again these rates are self-report — we're assuming it's valid because that's all we can do — less than 1 percent overall last year. The reason we think this is likely valid though is the lifetime rates are not as high as younger adult, but they're up there, which means they did disclose the event. It wasn't simply that they were afraid to talk about the event or embarrassed to talk about it, when we then did the follow-up questions regarding time frame, they said this happened when I was younger. So the sensitivity of the method appears to be there.

There was a significant difference by gender with sexual assault — as we expected, more women than men. Strangers again are not doing it — it's somebody in the person's environment, which means the risk is enduring. And again there's areas in potential intervention — maybe less so than the others. Social isolation of the perpetrator is a big problem, and that one's probably very difficult to address, but substance abuse, unemployment and mental health history are addressable problems. Their end was too low to do logistic regressions that we could feel good about the sample size in this one, so these are the univariate risk factors. We did do the logistics; I just don't want to report them — but social support came up high.

Now neglect. I define neglect in two ways — one in which there was an unmet need of living and there was a person identified as being able to take care of that need but not doing it. But really what I was interested in was potential neglect, or we could call it “societal neglect.” Here's an older adult that has an unmet need and nobody to help them take care of it. These are the areas — transportation, food and medicine; helping with cooking and preparing meals; helping with keeping the house clean and safe, and keeping the yard clean and safe; helping with getting out of bed, getting showered; and helping to make sure that financial things are taken care of — getting some money to pay the bills.

This one was much higher than we thought — these are people admitting they're having these types of problems. Again this is potential neglect, meaning they have a problem and there is no one to help take care of it. The big ones were maintenance of the house and yard and — well that was by far the biggest one — just general upkeep of where they live. There was again a gender split. Women had more problems with neglect. Low social support again — and this one did have a demographic set up. Being non-white and having low income increased neglect. Now that's logical; we probably didn't need a big study to do that. The low social support was again clearly logical in this mistreatment type.

Financial exploitation — this is the measure that I feel least good about because one, the time frame was very hard to get and two, it's a moving target of what's appropriate and what's not appropriate. You'll have to defer to Ken's [Conrad] stuff on this one because I feel that we did this not as well as the others. Areas: permission, good decisions — did they tell you about it and give you copies, forgery in the past, forced to trick you into signing things, stolen money. And then we switched to stranger — that was all familiar — stranger spending, stealing, forging, forcing you. Here's what we got — about 5 percent overall. That's within the past year by a family member. So 5 percent of the people felt they were being financially exploited by a family member within the past year. There was no real significant difference between men and women. This one, social support didn't come up with; it did in terms of stranger exploitation. Use of social services and ADL — so basically these were people who had some degree of functional impairment. They were exploitable for example — easier to take advantage of. And then this is the stranger — I'm going to go quickly through this one because this is a lifetime measure, so it's less relevant for older adults.

OK, to summarize ... If you look at any type of elder mistreatment in the past year — any at all — you'd get 15 percent not counting stranger financial abuse. Overlaying the most recent study (they did 3 of the 5 that we did), they had much higher rates of verbal abuse but much lower physical and financial. Their verbal abuse was very liberally defined. I thought ours was, but theirs was really liberally defined. I attribute our eight times higher rate of physical abuse to our methodology — it's been refined over about 10,000 people in different age groups, and I think we get that right. Here's your risk factor sets. The one in red is social support, so the abuse type is at the top and then the risk factors are down the side, and you can see some recurring themes — lower age, social support.

If you didn't notice I want you to focus on social support — and here's why. So basically you have 1 in 10 cognitively intact community-residing older adults getting one of the typical types of mistreatment — physical, sexual, emotional or neglectful; when you consider financial, it goes to 1 in 7. Low social support's the big player in the room. Now what do the perpetrators look like? They have relatively high unemployment — more than twice the national average. They have three to five times the national average of substance abuse. They're isolated, and they may have untreated mental health problems. These are targets that can be intervened. Neglect, low income, poor health and low social support — it didn't take a study to get that one. However, that, in combination with the idea of the social support as a risk factor across the board, tells us that social connection is going to be very important for older adults. Facilitating reconnection to the neighborhood, facilitating transportation — which is probably the single biggest problem in older adults — may actually reduce elder mistreatment. Social support is key. In my disaster research the most protective factor against developing psychopathology following disaster in older adults is social support. And this is easy — this can be something as simple as transportation, expanded Meals On Wheels programs, teaching communities to include older adults in most community-based initiatives, and things like that.

I think my time is up. Thank you.

[Applause.]