Caretakers do, in fact, report their abusive behaviors. Coyne et al. (1993) reported that 12 percent of caregivers calling a dementia care hot line indicated that they had abused the individual under their care. Homer and Gilleard (1990) studied respite care patients and caregivers in England and found that 45 percent of caregivers admitted either verbal (41 percent) or physical (14 percent) abuse. Interestingly, frequency of patient reports of abuse was less than that of caregivers. Similarly, Pillemer and Suitor (1992) interviewed family members of dementia patients and found that 6 percent reported violence. Pillemer and Finkelhor (1988) interviewed proxies when older adults were unable to participate as respondents and found higher rates of abuse than in victim reports (although this group of proxies arguably represented elders at increased risk, and higher levels should be expected). These studies, and studies cited below, demonstrate that caregiver assessment may be an acceptable, albeit unidimensional, method of detecting elder abuse in the subset of abusers willing to disclose these behaviors. Sensitivity can be expected to be increased if provisions for anonymity are enhanced.
When using caregivers as the data source, researchers have either assessed abusive behaviors directly through interviews or screens, or assessed risk factors associated with perpetrating elder mistreatment. Risk factors include alcoholism, social isolation, psychopathology, low socioeconomic status, overdependence on the older adult, and inexperience or reluctance to provide care (Reis and Nahmiash, 1998). In addition, caregiver risk-factor assessment can be augmented by studying care-receiver risk factors, such as being older, female, isolated, aggressive, or provocative.
Reis and Nahmiash (1998) developed the 29-item (from an original 48 items) Indicators of Abuse (IOA) screen based on previous risk factor research (Kosberg, 1988) with both caregivers and receivers. Although this is a screen, it requires prior in-depth knowledge of caregiver and care-receiver characteristics obtained through interview. The items were selected based on their discriminant ability to detect elder mistreatment derived as part of the major health and social services assessment offered in a North American city. Using the 29 items of the IOA that discriminated abuse from nonabuse, sensitivity was about 85 percent and specificity was 99 percent. Approximately 70 cases were reexamined by a panel to assure criterion accuracy. Using these criterion references, 28 of the original 29 items (caregiver age was dropped) achieved a sensitivity of 78.4 percent and a specificity of 100 percent. Factor analyses failed to identify separate thematic problem areas. Notably, items such as needing help with activities of daily life or cognitive or physical impairment did not contribute to discriminant ability. The overall findings indicate that caregiver rather than care