by the National Institute of Justice, indicates that households receiving the intervention had an increased risk of subsequent mistreatment (Davis and Medina-Ariza, 2001). The investigators adapted a model they had previously used in a study of the effects of a coordinated team response to family violence (Davis and Taylor, 1997). In the subsequent elder mistreatment study, the target population was persons living in selected public housing units in New York City who reported elder abuse (defined as physical abuse and psychological abuse) incidents to the police. Random assignment for the intervention occurred at two levels. First, 30 of 60 public housing projects were randomly assigned to receive public education about elder abuse (e.g., posters, leaflets, and project staff presentations). Second, in all 60 housing projects, half of the households reporting elder abuse incidents to the police were randomly assigned to receive home visits by a team of a police officer and a domestic violence counselor. The team discussed legal options and police procedures and attempted to link the households to social services. Victims were also encouraged to call the police if repeat incidents occurred. To determine whether abuse continued, police records were checked and victims were interviewed 6 and 12 months after the triggering incident.

Six months after the intervention, households receiving the home visit called the police significantly more often than controls, both in housing projects that received public education and those that did not. This is not surprising, since the home visit was designed to invite such reports. But this expectation was based on the assumption that the intervention would change reporting behavior, not that it would increase incidents of abuse. (If anything, one might have expected the number of actual incidents of abuse to be reduced due to deterrence.) The surprising finding was that that the increased number of calls was accompanied by an increased number of incidents of abuse, as reported by the victims to the research interviewers. That is, when households received both home visits and public education, victims of elder abuse reported significantly higher levels of physical abuse than households that received neither intervention or only one of them. During the period between 6 and 12 months after the intervention, the differences in calls to the police disappeared, but households that received the dual intervention continued to report significantly more incidents of physical abuse to the interviewers.

The researchers have speculated about the possible explanations for this paradoxical finding, including the possibility that the intervention angered the perpetrators. (As they pointed out, however, the perpetrators were not interviewed.) The most pertinent observation from the panel’s perspective is that the study raises more questions than it answers. Even well-intentioned interventions may have unexpected, and even harmful, outcomes.



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