ment events emerge. Still, the use of supplementary elder mistreatment modules within existing or planned large-scale or national surveys would seem to be a potentially fruitful approach that should be further evaluated.
The issue of household sampling is paramount in defining elder mistreatment occurrence rates. The general experience of household surveys targeting elders is that the older and more vulnerable potential respondents are the ones most difficult to access, leading to the potential for underassessment of elder mistreatment occurrence. Thus, in many instances, supplementary sampling approaches may be needed, such as through informal social networks, the health care system, or other social institutions.
The limitations of interview data as the sole source of elder mistreatment occurrence are apparent. Some of those at greatest risk, as noted above, may not be able or willing to serve as survey respondents, and while Acierno (this volume) notes that family members and others may admit to elder mistreatment, the completeness and accuracy of such declarations are uncertain. Inaccurate recall among survey respondents in general and older respondents in particular is well described, and recall accuracy is further called into question by the increasing levels of cognitive impairment with advancing age. In fact, Acierno (this volume) begins his discussion of case detection methods by dichotomizing elder mistreatment victims into those with and without “significant” cognitive impairment. This may be a useful construct, but cognitive function is multidimensional and variably progressive, so from the perspective of studying elder mistreatment occurrence, it may not be easy to categorize case populations into those with and without cognitive decline in advance of applying the case ascertainment protocol itself. Clearly, the issue of determining instances of elder mistreatment among those cognitively impaired is an important research question.
One potentially important method for enhancing knowledge of the occurrence and clinical course of elder mistreatment is record linkage. Determining the health, social, and economic status of older persons may profitably be enhanced by compiling information from many sources, including information from prior surveys, vital records, health care and health administrative records, social service and criminal justice records, and records from other publicly available, potentially health-relevant sectors of society. The use of primary institutional records should increase the accuracy of the information available for analysis and could complement information gained from interviews. However, there are several potential challenges to record linkage, including additional costs, the availability of electronic record systems, increasing privacy concerns (National Research Council, 2000a), and the logistics of assembling data from multiple sources.