Elder mistreatment is a recognized social problem of uncertain, though probably increasing, magnitude. Based on the best available estimates, between 1 and 2 million Americans 65 or older have been injured, exploited, or otherwise mistreated by someone on whom they depended for care or protection (Pillemer and Finkelhor, 1988; Pavlik et al., 2001). The number of cases of elder mistreatment will undoubtedly increase over the next several decades, as the population ages. Yet little is known about its characteristics, causes, or consequences or about effective means of prevention or management. This report is meant to point the way toward better understanding of the nature and scope of the problem, a necessary condition for the development of informed policies and programs. After summarizing the social context within which the field has developed, this chapter assesses the present state of knowledge, identifies some of the problems that must be addressed if the field is to move forward, and locates the problem of elder mistreatment in a larger set of challenges confronting an aging society.
AN AGING AND VULNERABLE POPULATION
The aging of the population of the United States is a well-recognized demographic fact. The life expectancy of people born in the United States has been rising throughout the past century. The proportion of the population age 65 and older has increased dramatically since 1950. Between 1950 and 2000, the total population of the country increased by 87 percent, the population age 65 and older increased by 188 percent, and the population 85 and older increased by 635 percent (Eberhardt et al., 2001, Hetzel and Smith, 2001). Over this same period, the life expectancy of people at age 65 increased from 13.9 to 17.9 years (Natonal Center for Health Statistics, unpublished data, 2001). These trends will likely be accentuated by the aging of the post-WWII baby boom generation. The U.S. Bureau of the Census predicts that by 2030, the population over age 65 will nearly triple to more than 70 million people, and older people will make up more than 20 percent of the population (up from 12.3 percent in 1990) (Population Projections Program, 2000).
It is heartening that large proportions of the nation's older people are living without substantial disability. Among people age 75 and older in 1999, 70 percent described their health as good or excellent (Eberhardt et al., 2001). Inevitably, however, the aging of the population is also associated with increases in age-related diseases and disabilities. Of the estimated 12.8 million Americans reporting need for assistance with activities of daily living (ADLs—eating, dressing, bathing, transferring between the bed and a chair, toileting, controlling bladder and bowel) or instrumental activities of daily living (IADLs—preparing meals, performing housework, taking drugs, going on errands, managing finances, using a telephone), 57 percent (7.3 million people) were over the age of 65 (Administration on Aging, 1997). Dementia is present in approximately 5 to 10 percent of persons age 65 and older and 30 to 39 percent of persons age 85 and older (Rice et al., 2001; Henderson, 1998). Among people age 85 and older in 1999, 33 percent reported themselves to be in fair or poor health, 84 percent had disabilities involving mobility (unpublished data Natonal Center for Health Statistics, 2002), and 16 percent had Alzheimer's disease (Brookmeyer et al., 1998).
Given the projected growth in the elderly population, long-term care for elderly people with disabilities has become an increasingly urgent policy concern (Institute of Medicine, 2001; Stone, 2000). The settings in which long-term care is provided depend on a variety of factors, including the older person's needs and preferences, the availability of informal support, and the source of reimbursement for care. An increasing number of elderly people reside outside traditional home settings in highly restrictive institutional environments (such as skilled or intermediate nursing facilities) or in less restrictive community-based residential settings, such as assisted living facilities, board and care homes, and adult foster homes. Among the 34 million persons over age 65 in 1995, 5 percent were nursing home residents, and 12 percent lived in the community setting with ADL or IADL limitations. The number of nursing home residents increased between 1973–1974 and 1999 from 961,500 to 1,469,500 among those age 65 and older, and from 413,6000 to 757,100 among those 85 and older (Eberhardt et al., 2001). In 1999, another 500,000 elderly people were living in assisted living facilities (Hawes et al., 1999). Among people age 85 and over, 21 percent were in nursing homes in 1995 and 49 percent were community residents with long-term care needs (Alecxih et al., 1997).
The nursing home population tends to be older and more severely disabled than elders residing elsewhere, with about half of the residents being 85 or older and about half having five ADL limitations, in 1996 (Stone, 2000); still, four out of five elderly persons with ADL or IADL impairments lived in the community setting (Alecxih et al., 1997). Approximately 17 percent of these community-dwelling older persons are considered severely disabled, with limitations in three or more ADLs. Of those ADL-impaired elderly people living in community settings, 37 percent report that they need help but do not receive it or receive less help than is needed (Stone, 2000).
Most long-term care for community-dwelling elders is provided in a traditional home setting, either in an older person's own home, with or without a spouse, or in the home of a close relative. The 1994 National Long Term Care Survey indicated that more than 7 million Americans, mainly family members, provided 120 million hours of care to elders with functional disabilities living in the community. However, the nature and character of the informal networks now providing long-term care services may change (Stone, 2000). The potential pool of adult children who can serve as caregivers is already decreasing, as a result of a variety of demographic trends, including divorce, smaller families, and increased workforce participation (Himes et al., 1996). These factors increase the pressures on families caring for their elderly relatives and also are likely to increase the demand for institutional care.
These trends highlight the growing challenge of ensuring the safety and protecting the other interests of elderly people in the diverse settings in which long-term care is provided. No matter where they reside, older people are vulnerable not only to the infirmities and suffering associated with disease and disability, but also to neglect, victimization, and exploitation by others, including their caregivers. In this respect, protecting older people from mistreatment is an important element of the broad challenge of ensuring quality services in long-term care.
While elder mistreatment has attracted sustained efforts from practitioners and some interest from policy makers over the past two decades, it has not received concomitant attention from researchers or from the agencies that provide research funding. No major foundation has identified this field as one of its priorities, and the federal investment has been modest at best. For example, fewer than 15 studies on elder mistreatment have been funded by the National Institute on Aging (NIA) since 1990, and support from other agencies has been even less substantial. As a result, elder mistreatment research has thus far been confined to a small community of investigators who have produced a modest body of knowledge concerning the phenomenology, magnitude, etiology, and consequences of elder mistreatment. Preventive and remedial interventions have been unsystematic, episodic, and poorly evaluated. In recognition of these deficiencies, the National Institute on Aging requested the National Research Council to commission this study as the first step in an effort to broaden and deepen knowledge about the mistreatment of elders. Support was also provided by the Office of Behavioral and Social Science Research on Women's Health of the National Institutes of Health and the Agency for Health Care Research and Quality. This report presents a research agenda for consideration by the National Institute on Aging and other potential sponsors of research on elder mistreatment—a term we explain more fully in Chapter 2.
HISTORICAL AND SOCIAL CONTEXT
Research on elder mistreatment is in an early stage, reflecting its relatively recent recognition as a distinct—and important—social problem. The prevailing understanding of the problem, and the social response to it, have gradually emerged over the past half-century, shaped by evolving social responses to child protection and family violence as well as by an intensifying concern about neglect and victimization of vulnerable elderly people.
Family discord and mistreatment of its vulnerable members were outside the public domain for much of this country's history. Responsibility for assisting families in need was assumed mainly by religious organizations and private charitable institutions. Although many states established asylums for people with mental illness during the 18th and 19th centuries, thereby providing some custodial protection for dependent or neglected adults, there was no legal basis for intervention into families until the late 19th century, when industrialization, immigration, and urbanization exacerbated family problems, including poverty and internal conflict, and also exposed them to public view—especially when its victims were children. The emergence of the juvenile court in the early part of the 20th century represented a significant assertion of collective responsibility for protecting and “saving” children who had become ungovernable by their parents; over the following decades, the jurisdiction of the juvenile courts gradually reached children who were neglected or abused by their parents (Platt, 1969).
The legal foundation for modern policies and programs for elder protection was put in place after World War II, particularly during a burst of national energy geared toward remediation of endemic social problems during the 1970s. Although the threads of child protection, adult protection, and family violence were intertwined in the history of that period, they are summarized separately below.
Origins of Child Protection
The current system for protection of elders and other vulnerable adults grew from the child protection system, which itself is only about 40 years old in its modern form. The seminal event in the formation of the modern child protection system was the publication of an article in the Journal of the American Medical Association by a team of physicians at the University of Colorado, who proclaimed the existence of a “battered child syndrome” (Kempe et al., 1962). Pediatrician Henry Kempe, the leader of the group and founder of the International Society for Prevention of Child Abuse and Neglect, spearheaded a movement to adopt mandated reporting laws. These laws, which were quickly adopted in all 50 states, rested on the premise that the abused child was an aberrant problem (amounting to several hundred egregious cases each year in the United States), and on the belief that the problem could be solved if health professionals brought those cases to the attention of social service authorities. Although initial federal action did not occur until significantly later, with the adoption of the Child Abuse Prevention and Treatment Act of 1974 (Nelson, 1984), that legislation also required states to adopt mandated reporting and investigation as the primary strategy for protecting children.
Origins of Adult Protection
Drawing on their parens patriae authority to protect helpless citizens, a few states developed new public welfare programs during the 1940s and 1950s to protect adults who could not manage their own resources or protect themselves from harm. New adult protective services units were established not only to provide social services, but also to provide legal services, such as guardianship. Aroused by these state innovations, federal interest in the problem first appeared in the 1960s. Legislation was directed at all adults who were seen as defenseless and susceptible to being hurt by others. In 1962 Congress passed the Public Welfare Amendments to the Social Security Act, authorizing payments to the states to establish protective services for “persons with physical and/or mental limitations, who were unable to manage their own affairs … or who were neglected or exploited” (U.S. Department of Health, Education, and Welfare, 1966).
One of the demonstration projects funded by this new program was operated by a team at the Benjamin Rose Institute in Cleveland under Margaret Blenkner and her associates (Blenkner et al., 1974; Anetzberger et al., 2000). She matched a group of elders receiving protective services with a group from the community who were receiving traditional services, finding that those who were receiving protective services had a higher mortality rate and higher nursing home placement rate than those who were receiving traditional services. This study raised important questions: Was the higher risk attributable to the intervention or to selection bias, and if the former, what aspect of the intervention increased the risk? Was it the nursing home placement? Notwithstanding this puzzling finding from the Blenkner study and other studies questioning the cost-effectiveness of protective services (Wolfe, this volume), advocates for the system continued to press for broader congressional action. Eventually, in 1974, Congress amended the Social Security Act to require states to establish protective service units for adults with mental and physical impairments, who are unable to manage on their own, and who were victims or were being exploited or neglected. Funding for the protective services was to come from social services block grants (SSBG) given by the federal government to the states. Until this time, most SSBG funds had been used exclusively for child protective services.
This new federal program directed the states to provide protective services to adults who, “as a result of physical or mental limitations, are unable to act in their own behalf; are seriously limited in the management of their affairs; are neglected or exploited; or are living in unsafe or hazardous conditions.” A number of states then codified this federal mandate and, by 1978, 20 states had legislation establishing adult protection units as part of their social services agencies. This trend was accompanied by increasing use of SSBG dollars for adult protection: in 1980, 38 states reported that 83.3 million SSBG dollars were spent for adult protective services. As SSBG appropriations declined during the 1980s, however, funding for adult protective services declined; by 1985, it had declined by 42 percent.
Spotlight on Elder Protection
Scarce attention was paid to the problem of elder abuse before 1978 except for some intermittent articles published in British and American medical and social services journals. In the late 1970s, the national spotlight was directed for the first time at what was characterized as systematic mistreatment of elderly people. Congressman Claude Pepper held widely publicized hearings, calling attention to the “hidden problem” of elder abuse in the nation's families, including what one witness characterized as “granny battering” (Wolfe, this volume). Although the Pepper hearings did not lead immediately to federal action or funding, they stimulated additional state action. As the state response continued to evolve in the early 1980s, many states required reporting of abuse, bringing the problem within the purview of adult protective services. By 1985, 46 states had designated a responsible agency. Meanwhile, Congressman Pepper continued to agitate for a federal response to elder mistreatment. In a 1981 report (Pepper and Okar, 1981), he stated that elder abuse was increasing and recommended that Congress act immediately to help the states identify and assist elder abuse victims. Again, however, Pepper's plea was unheeded by the Congress. Finally, in 1989, Pepper succeeded in including creation of a national center on elder abuse as an amendment to the Older Americans Act. Although various versions of a national center followed, the current National Center on Elder Abuse was established in 1998.
In retrospect, it appears that elder mistreatment became identified as a national concern when it was conceptualized as an “aging” issue, rather than as an undifferentiated component of adult protection. This also helped to broaden the constituencies interested in research and program development to include gerontologists and the expanding network of service providers and advocates for the elderly. The Pepper hearings also cast the problem of elder abuse in a particular light—as a complication of caregiving. The emerging image was that of an impaired victim, usually an elderly parent being cared for by an adult caregiver who wasn't able to manage the caregiving because of stresses in life, on the job, and in the family. Even though it is only a partial explanation of elder mistreatment, this picture seemed to resonate with Congress and the media (Wolfe, this volume).
Emerging Conceptions of Family Violence
The evolving understanding of elder mistreatment as a social problem has more recently been shaped by another image—the trapped victim of family violence. Spouse abuse and other varieties of intimate partner violence have received increasing professional and political attention since the 1980s, leading to a wide variety of interventions and a substantial investment in research (National Research Council, 1996; National Research Council and Institute of Medicine, 1998). Prevention, protection, and punishment are necessary components of a comprehensive social response, requiring the participation and coordination of a broad array of public agencies. As the consciousness of health professionals has been raised, family violence has been embraced as a public health problem, thereby recruiting researchers and advocates in injury prevention and public health to the field (Institute of Medicine, 1999). Many of the preventive and protective tools developed in the context of intimate partner violence have now been directed to violence against elders. Bringing elder mistreatment into the domain of family violence widens the angle of the lens and thereby brings new ideas about etiology and prevention into view. However, it also exposes some tensions between social services agencies, with their traditional helping orientation, and many family violence specialists, with their greater emphasis on criminalization and punishment of perpetrators.
The Crisis in Child Protection
Evolving conceptions of elder mistreatment, and the appropriate social responses to it, will also be shaped, inevitably, by the deep concerns that have emerged over the past decade in the field of child protection. In 1990, the U.S. Advisory Board on Child Abuse and Neglect issued a highly publicized and rarely disputed declaration of a national emergency in the child protection system. By that time, the number of cases reported annually to state and county social service and law enforcement agencies in the United States approached 3 million—a number enormously discrepant from the 1962 estimate of Kempe et al. of approximately 300 cases annually. Moreover, the advisory board found that, by state social service agencies' own admission, many children officially found to have been maltreated received no services other than the investigation itself.
The U.S. Advisory Board on Child Abuse and Neglect (1990) attributed the emergency to the errant design of the child protection system itself: the system has become preoccupied by investigation (rather than prevention and treatment), and community responsibility for ensuring the safety of dependent children has effectively, if unintentionally, been diverted to a small social service agency. In response, the board (U.S. Advisory Board on Child Abuse and Neglect, 1993) proposed a new national strategy designed to rely on voluntary action to make child protection a part of everyday life (see Melton and Barry, 1994, Melton et al., 2001, for edited books articulating the social science foundation for this approach). As Wolfe notes in his paper in this volume, several states have attempted to deemphasize investigation in their state child protection statutes, and some major foundations have undertaken initiatives to demonstrate the feasibility of a neighborhood-based, largely voluntary, and largely preventive and supportive child protection system. Nonetheless, modal practice is largely unchanged, and the enormity of the problem remains (Melton, 2002).
The tensions in child protection policy (as well as the number of reported cases) have intensified as the scope of problems defined as child maltreatment has expanded. Although the modern system was created in response to the image of battered children, neglect has long been the modal reason for referral to child protection (Peddle and Wang, 2001), and most such cases involve complex social and economic problems, not willful neglect (Pelton, 1994). Similarly, the biggest increase in reporting occurred when sexual abuse was “discovered” early in the 1980s (Weisberg, 1984), and criminal prosecution became a common feature in the child protection system.
Recognition of the frequent linkage between intimate partner violence and child maltreatment (see Carter et al., 1999) has also challenged the child protection system, which generally (except to some degree in cases of sexual abuse and severe physical abuse) has not adopted the “perpetrator-victim” model commonly embraced by advocates for battered women (Melton and Andrews, 2000). There are some signs of an uneasy rapprochement between the two systems (see, e.g., Schechter and Edleson, 1999), as some child protection authorities have adopted safety planning, a feature of victim empowerment in programs for battered women, as a potentially useful element of intervention in cases of child maltreatment.
Even this development, however, has illustrated the field's vulnerability to unintended side effects. For example, a legislative determination in Minnesota that exposure of children to intimate partner violence is per se evidence of child neglect led to an immediate doubling of referrals to child protective services, a huge increase in expenditures, and increased stress and loss of confidentiality for women and their families living in shelters (Edleson, 2000). It was also speculated that this policy, soon retracted by the legislature, deterred some battered women from seeking protection for themselves and their children.
These tensions and policy adaptations in the field of child protection appear to be highly relevant to elder protection at this moment in the evolution of research and public policy in this nascent field. As discussed further in Chapter 6, adult protection services agencies grapple daily with the tensions between investigation and service, and prosecution and protection. Agency caseloads reflect the highly diverse problems within their jurisdictions, ranging from intentional partner violence to far more numerous cases of caregiver neglect (as well as problems not arising in child protection, such as financial exploitation). The recent history of child protection offers many lessons for specialists in elder mistreatment.
Prevailing conceptions of elder mistreatment draw on a diverse array of images (the forgotten and helpless nursing home resident, the battered granny, the stressed caregiver, the abusing spouse). Moreover, the system of adult protection that has emerged to respond to these varied problems (as well as other problems relating to adults with disabilities) is based on ideas and structures borrowed from policy and practice in child maltreatment and, more recently, intimate partner violence. Yet prevailing policies and practices in these adjacent domains are not fully applicable to elder mistreatment and have been controversial on their own terms. Repeatedly, National Research Council and Institute of Medicine panels have called attention to the need for sustained and aggressive research on the phenomenology, magnitude, etiology, and consequences of these problems and on the effects of interventions (National Research Council, 1993, 1996; National Research Council and Institute of Medicine, 1998). In so doing, they have noted that very little is known about the phenomenology, magnitude, etiology, and consequences of elder mistreatment, and that almost nothing is known about the effects of interventions. Although the body of evidence remains sparse, researchers have recently begun to raise doubts about the cost-effectiveness of current interventions (Dyer et al., 1999; Harrell et al., 2002; Pavlik et al., 2001; Hajjar and Duthie, 2001; Wolf and Li, 1999).
Overall, the national response to elder mistreatment still remains weak and incomplete. Adult protection is a poorly funded system, and Congressman Pepper's single-minded emphasis on the abuse, exploitation, and neglect of vulnerable elderly people has not been sustained by his successors in Congress or by a public preoccupied with youthfulness and ill at ease with aging. As a result, elder mistreatment remains hidden, poorly characterized, and largely unaddressed—more than two decades after the Pepper hearings first exposed it to public view. It is long past time to move the field forward in a careful and systematic way, drawing on the knowledge already generated in the domains of child maltreatment and intimate partner violence, while remedying the weaknesses that have so far plagued the field.
WEAKNESSES IN EXISTING RESEARCH
Although there is a sizable body of unpublished reports and commentary on elder mistreatment, fewer than 50 peer-reviewed articles based on empirical research have been published in the field. (A summary of these studies appears in Appendix A.) Although these studies provide a foundation for further work, it is not a strong one. National Research Council (1993) and Institute of Medicine reports (2001; National Research Council and Institute of Medicine, 1998) and other authoritative reviews (e.g., Pillemer, 2001; National Institute of Justice, 2000) have repeatedly lamented the weakness of the research base for designing programs and informing policy on the wide variety of overlapping problems, ranging from granny battering to neglect by nursing homes, that are grouped under the rubric of elder mistreatment. A systematic program of research is needed to better describe the many facets of the problem and to explore their causes and consequences.
Understanding the nature and scope of the problem is prerequisite to designing and implementing solutions. In the absence of the necessary research, interventions have been designed and implemented in the dark, so to speak. Almost every state has required reporting of suspected cases of elder mistreatment, but little is known about the effects of these requirements (National Research Council and Institute of Medicine, 1998). A few states and localities have mounted some creative interventions, but these few initiatives have been poorly evaluated. It has often been said that elder mistreatment, as a field of research, is at about the same embryonic stage of development as child mistreatment was about 30-40 years ago.
Some of the weaknesses of elder mistreatment research are summarized below.
Unclear and Inconsistent Definitions
The first major difficulty in analyzing results from previous research on elder abuse and neglect results from the poor definition of the term “elder abuse.” To some extent, this problem is a reflection of conceptual confusion: What type of behavior or condition is denoted by the concept of “abuse”? To some extent, it is also traceable to the variations and ambiguities of the state statutes that direct or authorize interventions in cases of elder abuse or neglect. (The statutes are discussed in Chapter 2.) However, researchers have often exacerbated the problem by failing to define or operationalize their terms in a clear and objective way. For example, many researchers refer to the entire range of problems experienced by elders as “abuse,” including lack of proper housing, untreated medical conditions, and lack of social services. Most of the studies are weakened by their undifferentiated treatment of various types of abuse and neglect. That is, all forms of mistreatment are lumped together, despite evidence that the forms of abuse and neglect differ substantially. In some studies, for example, it is difficult to determine whether financial exploitation is included in the research definition. Studies are especially weakened by their inclusion of the category “self-abuse” or “self-neglect.” As discussed below, these terms refer to a category of conditions that has little in common with the conditions that bear on abuse and neglect of elder persons by other people.
Researchers have also diverged widely in their definitions of the pertinent component terms and have frequently used confusing and unclear definitions. For example, some researchers have used the term “abuse” tautologically; for example, one group of researchers defined elder abuse as “an abusive action inflicted by the abusers on adults 60 years of age or older.” Another group called elder neglect and abuse “a generic term that refers to the neglect and/or physical, psychological, or financial abuse of the older person.” Furthermore, definitions have differed so widely from study to study that the results of research are almost impossible to compare. While one set of investigators calls “withholding of personal care” physical abuse, a second researcher calls it active neglect; a third subsumes such actions under physical neglect; and yet a fourth considers such behaviors to be “psychological neglect.” Similarly, some researchers define physical abuse in terms of actions: hitting, pushing, choking, etc. Others, however, use lists of injuries to define physical elder abuse, such as cuts, fractures, bruises, and burns.
The development of better definitions of mistreatment of the elderly should be an extremely high priority for researchers. In particular, it is critical to differentiate among various types of mistreatment. Researchers must be clear and explicit regarding what is included and excluded from the category of elder abuse in order to conduct any meaningful meta-analyses. The panel addresses this problem in the next chapter.
Unclear and Inadequate Measures
Related to the definitional issue is that of measurement. This is an equally vexing problem, since the definitions of the varying elements of elder abuse must be operationalized through the design and administration of a research instrument. Many studies have not developed separate research instruments at all; instead, they have simply analyzed the forms used by agencies. These forms are not designed for research and rarely provide data of the type and quality to be of use to researchers. Or studies use as a “measure” of abuse whether a professional has identified an elderly person as “abused”—thereby embracing without further clarification the discretionary judgments of clinicians and caseworkers applying the ambiguous statutory definitions. Few attempts have been made to create reliable and valid instruments for the studies. Even when research instruments have been used, researchers have used highly varying approaches.
An example to illustrate this point may be in order. Researcher A includes physical abuse in her definition of elder abuse. She is using the Conflict Tactics Scale, which measures physical acting out in response to conflict. She then proceeds to define physical abuse as a single incident in which the elder is hit, bit, punched, kicked, threatened with a weapon, or has a weapon used on him or her. Researcher B also includes physical abuse in his definition of elder abuse. However, he has developed his own scale, similar to the Conflict Tactics Scale but more broadly constructed, so that it measures any assaultive behavior of hitting, biting, kicking, punching, threatening with a weapon, or using a weapon regardless of the reason for the behavior. Furthermore, he decides that there must be at least two episodes of this behavior for it to be called physical abuse except for those items dealing with weapons, in which case one incident is sufficient. Thus both researchers have included physical abuse in their studies—indeed, it may be the sole focus of each researcher's study—but the measure of physical abuse differs across the two studies.
This problem arises for all of the types of elder mistreatment typically investigated, including neglect and financial exploitation. The lack of definitional consistency poses issues for interpretation and understanding across studies, including determining prevalence and risk factors. However, even if researchers embraced a common set of definitions for the elements of elder mistreatment and operationalized them the same way, that would still leave the problem of determining whether the instruments actually measure what they purport to measure (validity) and whether they can be reliably administered. At the present time, no measure of elder mistreatment has been validated, nor has any instrument been embraced by the field as a definitive measure of mistreatment, even within a narrow sphere.
All this suggests that researchers, policy makers, and other consumers of research on elder mistreatment must pay careful attention to the definitions and measures of any studies on which they rely. In most cases, the measures will not be comparable.
Incompleteness of Professional Accounts
Since the earliest stages of elder abuse research, surveys of professionals have been used to shed light on the prevalence of elder abuse and on risk factors. Investigators typically mail surveys to professionals and paraprofessionals, asking them about contacts with cases of elder abuse or neglect during a given time period. To provide a typical example, in a survey on elder abuse funded by the Administration on Aging, a sample of professionals, including administrators and direct service workers from 16 types of agencies, was surveyed in each of Pennsylvania's 67 counties. Overall, one-half of the responding agencies reported encountering elder abuse, ranging from over 90 percent of domestic violence agencies, to less than 30 percent for law enforcement, emergency services, medical clinics, and drug/alcohol agencies (Fiegener et al., 1989). Similarly, a survey of Alabama physicians and registered and licensed practical nurses found that 38 percent of the physicians and 53 percent of the nurses had seen cases of elder abuse in the previous year (Clark-Daniels et al., 1990).
At best, studies of professional experience provide impressionistic estimates and opinions about the prevalence, correlates, and consequences of elder mistreatment. Although such data may be useful for generating hypotheses for further research, they do not provide a sound basis for designing programs or formulating policies.
Elder mistreatment researchers have also relied on samples of cases that have come to the formal attention of a social agency or reporting authority. For example, records of patients at hospitals or social service agencies have been reviewed, and the percentage of elderly persons judged to have been abused is established. A more controlled version of this kind of study provides agency caseworkers or health professionals with a standardized assessment tool, which they are trained to fill out for clients. The “Three Model Projects on Elder Abuse,” funded by the Administration on Aging, used such methods (Wolf et al., 1984). In both types of studies, however, researchers obtained data from professional accounts of mistreatment rather than from interviews with victims themselves.
It is widely recognized that reported cases are highly selective samples, and that there is a large reservoir of unreported and undetected cases of elder mistreatment about which very little is known. Although unreported cases may be similar to reported cases, they also may be quite different. Samples of reported cases may suggest common patterns and correlates of mistreatment, especially when paired with a control group, but the data must be interpreted with great care. Most important, the question of the extent of elder mistreatment cannot be answered by studies of reported cases. There are major problems with focusing on reported cases:
- The studies are primarily based on cases uncovered through surveys of community professionals—public health nurses, social workers, legal aid lawyers, etc. They are thus cases that have come to public attention in one way or another. However, we know from other studies of family violence using nonclinical populations that only a fraction of cases involving serious mistreatment comes to public attention and that these cases are not necessarily representative of the problem at large. (In relation to child abuse, for example, see the 1995 Gallup Poll, finding that far more of America's children are victims of physical and sexual abuse than officially reported—Gallup Poll, 1995.)
- Similarly, in most cases, the research data on elder mistreatment have not come directly from victims, but instead from professionals and outside observers. Such secondhand knowledge may distort the actual dynamics of mistreatment by failing to present the problems and their effects, as the actual participants perceive them.
- Case reports have little value in studying some forms of mistreatment that are rarely reported to adult protective services agencies, such as mistreatment in institutional settings.
Because elder mistreatment studies have relied so heavily on reports from professionals, crucial data about abuse situations have been missed. Community professionals in general do not collect data useful to researchers and policy makers. Thus, previous research using agency records has rarely been able to obtain detailed information about family history, attitudes, and consequences of mistreatment and other issues. Some researchers (e.g., Lachs et al., 1997a) have made effective use of these weak datasets by matching cases with higher-quality datasets.
In an effort to generate a national estimate of the occurrence of elder abuse and neglect based on case-identification by professional “sentinels,” the National Center of Elder Abuse, in conjunction with Westat, Inc., conducted the National Elder Abuse Incidence Study (National Center on Elder Abuse, 1998). In this study, modeled after recent incidence studies of child abuse, the researchers identified a nationally representative sample of 20 counties in 15 states; for each county sampled, they collected data from the local APS agency as well as approximately 1100 professional “sentinels” having frequent contact with the elderly. In 1996, according to the projections based on this study, about 450,000 persons age 60 or older experienced abuse or neglect in family settings, about 16 percent of whom were in the APS report files. It is generally acknowledged that these findings detect only the most overt cases and thus significantly underestimate the incidence of elder mistreatment.
Studies of professionals and agency records are justified in those situations in which investigators specifically want to know how professionals view elder mistreatment. But researchers have too often used these professional surveys to estimate the incidence or prevalence of elder mistreatment, or to establish its causes. They are not appropriate for these purposes. Future research in this area should go beyond archival data and should rely to a much greater extent on elder persons' accounts of their experiences and on their perceptions regarding their own security.
Lack of Population-Based Data
Data on the extent of elder mistreatment in the general population are sparse. Representative sample surveys of community populations are urgently needed. Over the past two decades, knowledge about violence in families and the victimization of children and other vulnerable people has improved significantly. A major advance has been the fielding of major population-based victimization surveys that have helped to establish reliable prevalence estimates of select problems, such as intimate partner violence and child physical and sexual abuse. Similar progress has not occurred in the field of elder mistreatment.
In the earliest research about two decades ago, studies were generally conducted on small, nonrandom samples, with little generalizabilty to the population. Furthermore, research in the field was conducted independently by investigators from different disciplines, using different methods and without recognizing the problems faced by other investigators. For example, the medical community focused on clinical signs and symptoms that could not be explained by disease markers, and this was a daunting task. Very often, older adults who had multiple chronic diseases or conditions might have symptoms that could mask or mimic mistreatment. Using a patient-based approach to study elder mistreatment is also fraught with potential for sample bias, in that if an older adult does not have a doctor or does not come to the emergency department, mistreatment cannot be evaluated.
Although some population surveys have subsequently been fielded, many of them have excluded from the sample potential respondents who may be at high risk for abuse or neglect—e.g., older adults with profound dementia, severe hearing or speech impediments, or advanced problems with mobility who are unable to participate in survey research. Although some investigators have tried to use proxy respondents, this method poses even more challenging issues, because the proxy may be implicating him or herself in mistreatment.
Prevalence information (for one community in the United States) was best established by Pillemer and Finkelhor (1988), who used a stratified random sample of community dwelling older persons (65 or older) in the Boston metropolitan area. A two-stage interview process was used: screening to determine if the person was a victim of mistreatment (defined to include physical abuse and psychological abuse and neglect but excluding financial abuse), followed by in-depth interviews by telephone or in person. Since 1988, there has been no effort in the United States to obtain better prevalence data using large-scale random samples on either a locally or nationally representative sample. However, four such studies have been undertaken in Canada (Podnieks, 1992), the United Kingdom (Ogg and Bennett, 1992), Finland (Kivela et al., 1992), and The Netherlands (Comijs et al., 1998). Despite using different methods, these studies each reported that the prevalence of elder abuse falls in the 3-5 percent range. (It should be noted, however, that the scope and content of the definitions used in these studies vary, particularly with regard to financial abuse.) Despite attempts to estimate incidence and prevalence in other ways, random sample surveys of the elderly population alone allow for a more accurate assessment of the rate of elder mistreatment. In the United States, a national survey is urgently needed to estimate the prevalence of different types of elder mistreatment in the general population, and in specific regions and subgroups, as well as the co-occurrence of different forms of mistreatment (see Chapter 4).
Lack of Prospective Data
Prospective studies are powerful designs, in that they can overcome the recall bias inherent in retrospective studies based on self-reported mistreatment. Studies of this kind are urgently needed: to date, no prospective study of elder abuse has been conducted. However, in a pioneering study, Lachs and colleagues retrospectively linked Adult Protective Services data to a prospective study—the New Haven EPESE study (Established Population for Epidemiologic Studies in the Elderly) as the basis for this research, one of four cohorts funded by NIA (Lachs et al., 1996). In inception year 1982, the study sample consisted of 2,812 community-dwelling older adults over age 65. A manual record matching of EPESE and Connecticut ombudsman/elderly protective service records was done to determine if any cohort members had been seen by ombudsmen over an 11-year follow-up period from cohort inception (1982-1992 inclusive). After cohort members who were seen by protective services for the elderly were identified, weighted survival curves from cohort inception were constructed for three subgroups of subjects: (1) those found to have sustained verified elder mistreatment (abuse, neglect, or exploitation) by another party (i.e., nonself-neglect), (2) those seen by protective services for corroborated self-neglect, or (3) other members of the cohort who had no contact with elderly protective services.
Lack of Control Groups
Much of the data on risk factors and consequences of elder mistreatment are drawn from studies of clinical case samples. However, few of these studies have used controlled designs. For this reason, generalizations made from the existing studies are necessarily suspect. For example, some investigators have asserted that the abused elderly tend to be physically or mentally impaired or both. However, without a comparison group, it is impossible to know if they are more or less impaired than other persons. Several studies have attempted to go beyond previous efforts by interviewing the victims themselves and including a control group of nonabused elderly persons (Bristowe and Collins, 1989; Paveza et al., 1992; Pillemer and Finkelhor, 1988). These are still few and far between, however. Interestingly, although a number of controlled studies were conducted in the late 1980s and early 1990s, there are virtually no examples of more recent case-control studies of elder mistreatment.
Lack of Systematic Evaluation Studies
There has been almost no effort to evaluate intervention programs for elder abuse. Certainly, no study has as yet attempted a randomized control group design in this area. Any kind of experimental demonstration project is rare. Little is known about the relative effectiveness of various programs.
Due to such shortcomings, existing studies have not provided adequate data needed to answer three important public policy questions about elder abuse and neglect:
First, is the problem of sufficient magnitude to warrant large-scale public concern, including such measures as mandatory reporting laws and protective services? Better data on the true prevalence of elder mistreatment are needed in deciding what action government ought to take.
Second, what are the characteristics of locations, conditions, situations, and relationships in which the elderly are most vulnerable to mistreatment? To design and implement intervention programs, policy makers and service providers must learn more about the factors that increase or decrease the risk of mistreatment and the conditions that ensure safety.
Third, what interventions prevent elder mistreatment and ameliorate its effects? Extensive evaluation research using scientifically sound research designs is critically needed.
IMPEDIMENTS TO ELDER MISTREATMENT RESEARCH
Why is knowledge about elder mistreatment so underdeveloped? What accounts for the paucity of sound research in this important area? The panel has identified a number of explanatory factors.
- Many investigators believe that victims and family members are not suitable respondents for interview studies of elder mistreatment, because they are not reliable respondents, because they are not willing to be interviewed, or because they are incapable of giving the necessary consent. In fact, many victims are more than willing to be interviewed and are reliable respondents able to give the necessary consent. Surveys including such respondents have uncovered serious cases of mistreatment, and a variety of studies have been conducted in which victims have been interviewed.
- In general, methods that have been used successfully to investigate other forms of family violence have not been applied to research on elder mistreatment. Gerontologists who study elder mistreatment have tended to follow their interests in family caregiving and have seen the problem in this context. However, because much elder mistreatment does not occur in family caregiving situations, this has been a serious limitation. Furthermore, the technology for studying family violence has been developed and refined not by gerontologists, but by child abuse and intimate partner researchers. Elder mistreatment researchers have not been trained in methods of studying other forms of family violence, including sampling methodologies and measurement techniques.One example of this problem is the lack of studies using the Conflict Tactics Scale (Straus, 1978; Straus and Gelles, 1990, Straus and Gelles, 1992) to study elder mistreatment. Regardless of the occasional controversy over the scale, it is a hallmark instrument that has been used in scores of studies of child abuse and intimate partner abuse. It is to some extent the state of the art, but some elder mistreatment researchers do not seem to be aware of it.
- It is very difficult to obtain access to perpetrators of mistreatment. In intimate partner studies, a number of researchers have used treatment programs for batterers as sources of research subjects. These do not exist for elder mistreatment.
- The exclusion of some victims can seriously bias samples. The problem is most evident when residents of institutions are excluded altogether from population samples. However, even within the targeted study population (whether community dwelling or residing in institutions), exclusion criteria based on cognitive deficiencies can seriously skew the findings.
- There is some anecdotal evidence that institutional review boards have interpreted the Common Rule (the governing regulations on research ethics) in an unduly restrictive fashion, impeding potentially valuable research on elder mistreatment (see Chapter 8).
- Few investigators have been drawn to this field of inquiry. Reviews of the literature reflect the same small set of names time and again, with few new researchers selecting and remaining in this field. One of the reasons for this situation is that so little funding has been available for research on elder mistreatment. Although more outstanding investigators might have attracted more funding, dedicated funding also could attract more and better investigators. Although the total federal contribution to research on elder mistreatment is uncertain, expenditures by NIA, the lead agency for aging research, have totaled $10 million during the last 12 years (1990–2001). Annual expenditures have increased from less than $300,000 per year in 1990 to over $1.3 million in 2001; this is a modest sum even in comparison to the underfunded domain of child abuse research, on which federal agencies spend $3.8 million each year.
- The existing body of research is largely descriptive and pragmatic, taking the concepts and definitions used in practice or in statutes as given, rather than deriving the concepts and measures from theoretical premises or hypotheses. The atheoretical nature of the research is reflected in the tendency to lump all forms of mistreatment within a single category.
- Individuals who have attempted to conduct research on elder abuse report that they have sometimes been hindered by a lack of cooperation from agencies responsible for identifying and treating victims of mistreatment. Adult protective services programs and other elder abuse service programs have been characteristically reluctant to assist researchers in research activities, and especially research that involves interviews with victims and their families. Reasons for lack of agency cooperation include a desire to protect their clients' privacy and to prevent additional disruption in their lives, fear of evaluation research, and a shortage of staff time to devote to research.
- Although every state has enacted a statute authorizing or directing intervention in cases involving vulnerable adults, including the elderly, these statutes vary widely in almost every respect (see Appendix B and tables in Chapter 2). They specify different ages or circumstance under which a victim is eligible for protective services, often differentiating between in-home and institutional abuse. They also vary in definitions of abuse, classification of abuse as civil or criminal, whether reporting is mandatory or voluntary, and the remedies or resources available when abuse is documented.
Each of the statutes defines conditions or circumstances that warrant intervention. The statutes typically define abuse or mistreatment as a series of broad categories, such as physical abuse, psychological or emotional abuse, sexual abuse or exploitation, and fiduciary abuse or exploitation, as well as neglect. However, not all states include all of these categories, and others are sometimes added. For example, some states do not include psychological abuse within the definition, while others add more specific forms of mistreatment such as “unreasonable confinement” or “abandonment.” Moreover, statutes sometimes distinguish between degrees of mistreatment according to the perpetrator's culpability or state of mind; for example, the law may distinguish among willful infliction of physical abuse, negligently causing physical injury, and failure to prevent it.
In addition to variations in the types of mistreatment included in the statutory definition, the statutes also differ substantially in defining the common categories. For example, the definition of emotional abuse in several states includes “ridiculing or demeaning an infirm adult, making derogatory remarks to an infirm adult or cursing or threatening to inflict physical or emotional harm on an infirm adult,” whereas other states require proof of “extreme emotional distress or harm” (see Appendix B).
These statutory variations in definitions and obligations create innumerable opportunities for confusion and lack of comparability, especially if reported cases are being studied. When data are reported to some central repository, unless the repository has imposed a specific definition for each of the forms of abuse, the same statutory element will trigger reports in different categories of cases in different states. Interpretation of combined statistics is treacherous, even if the only objective is to compare trends across states.
OUTLINE OF REPORT
Keeping in mind the impediments to research identified in this chapter, the panel decided to concentrate its attention on the tasks that are most urgently needed to propel the field forward. Chapter 2 addresses the problem of inconsistencies in definition and measurement that have thus far characterized research on elder mistreatment. Chapter 3 sketches a theoretical framework that may be useful in organizing research on the phenomenology and etiology of elder mistreatment in different settings and contexts. Chapter 4 addresses the challenge of measuring the occurrence of elder mistreatment in the population, highlighting important epidemiological considerations in elder mistreatment research. Chapter 5 summarizes what is now known about risk factors for elder mistreatment and identifies priorities for future research. Chapter 6 addresses research needed to improve screening and case identification in clinical settings. Chapter 7 reviews policies and programs aiming to prevent or respond to elder mistreatment and identifies priorities for future research. Chapter 8 addresses concerns about protecting human subjects in elder mistreatment research, and Chapter 9 identifies some necessary conditions for moving the field forward. The panel's conclusions and recommendations are presented in Table 1-1.
National Academies Press (US), Washington (DC)
National Research Council (US) Panel to Review Risk and Prevalence of Elder Abuse and Neglect; Bonnie RJ, Wallace RB, editors. Elder Mistreatment: Abuse, Neglect, and Exploitation in an Aging America. Washington (DC): National Academies Press (US); 2003. 1, Introduction.