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,



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement