To ascertain the extent of, and risk factors for, physical abuse among older urban American Indian/Alaska Natives (AI/ANs), we conducted a chart review of 550 urban AI/AN primary care patients Ն 50 years old seen during 1 year. Mistreatment was documented in 10%. A logistic regression found younger age ( P Ͻ .001), female gender ( P Ͻ .001), current depression ( P Ͻ .001), and dependence on others for food ( P Ͻ .05) to be significant correlates of physical abuse. In only 31% of instances of definite abuse were the authorities notified. We conclude that providers should be alert to the possibility of physical mistreatment among older urban AI/ANs. Improvements in detection and management are sorely needed. KEY WORDS Elder abuse is recognized as a national problem, yet little is known about victimization of American Indian or Alaska Native (AI/AN) elders. Rates of abuse reported among AI/ANs have ranged from more than 2% to 46%, with probable variation across tribes by socioeconomic factors. [1][2][3] Typically, these surveys have used convenience samples and not examined abuse in urban or health care settings. Additionally, among providers serving AI/ANs, lack of recognition may be combined with the mistaken assumption that, because of the respect Native cultures accord older adults, their mistreatment is unlikely. Thus, it is not surprising that official reports of abuse of AI/AN elders are rare. 4 Because our experience indicated otherwise, we reviewed the medical records of 550 older AI/ANs to examine the frequency of, and risk factors for, physical mistreatment. METHODSElders were defined as persons Ն 50 years old because in AI/AN culture, "elder" status is not solely a function of chronological age and may be conferred earlier than in the majority culture. Furthermore, impairments associated with aging in AI/ANs may occur 20 years earlier than in the general population. 5 Since elders constitute a much smaller percentage of the AI/AN population compared with other minorities, 6 lower age eligibility criteria also exist for some federally and state funded programs.This study was conducted at the Seattle Indian Health Board, which provides health care to the King County Native population. Overall, approximately 50% of their clients are unemployed; 80% have incomes below the "poverty line"; and 58% have no health insurance. The Seattle Indian Health Board's computerized information system was used to identify all AI/ANs Ն 50 years old seen between June 1994 and June 1995; this process yielded 550 names.The charts of all 550 patients were reviewed by a professional abstractor who was trained and supervised in the chart review procedures by a general internist (DB). The steps in this training included multiple sessions in which the study forms were developed, pilot tested on actual medical records, discussed with the general internist, and then modified. Subsequently, charts were reviewed specifically for discrepancies in categorization and completeness between the general internist and the abstractor. Because ...