Background. Elder abuse (EA) is the physical, sexual or psychological abuse, financial exploitation or neglect of an adult age ≥60 years. One in 10 older adults experience EA annually in the US, with many experiencing multiple types. Veterans are at particularly high risk due to the high prevalence of EA risk factors in this population. Experiencing EA is linked to depression, injury, increased healthcare use and mortality, but despite its prevalence and morbidity, fewer than 5% of cases are detected, limiting opportunities for intervention. While screening is a common approach to improving detection of similar conditions, screening tools for EA have not been well validated or widely studied. Furthermore, EA screening may miss important high-risk populations, such as those with dementia, necessitating the development of additional detection strategies that complement screening. This research aims to improve EA risk detection in VA by both evaluating and optimizing current EA screening approaches and by leveraging VA healthcare data to identify Veterans with clinical suspicion of EA who may benefit from further assessment. Significance/Impact. With the growing population of older adults in the US and over 10 million US Veterans age ≥60 years, improving detection of and interventions for EA is a national and VA public health priority. By improving detection of EA via both better-informed screening and novel data-driven tools, this research aligns with VA HSR&D’s priority to improve care for our nation’s aging Veterans and their caregivers. Innovation. This research integrates elder abuse and implementation science conceptual frameworks to develop new approaches to improving EA detection. This study will evaluate the test characteristics of the first-ever data marker for EA suspicion using unique VA data elements and will employ innovative data informatics approaches, such as natural language processing (NLP), to address a complex social problem with large health impacts. Specific Aims. Aim 1 is a national assessment of the current landscape of EA screening practices in VA medical centers (VAMCs) and a quantitative evaluation of facility level factors associated with screening. Aim 2 is a quantitative study that will identify the best performing EA administrative marker (AM) in VA data. Aim 3 is a qualitative study that will elucidate opportunities for, facilitators of and barriers to implementation of healthcare-based EA detection programs in VA. Methodology. In close partnership with the VA Office of Care Management and Social Work, Aim 1 will conduct a national survey of VAMCs to assess current practices around EA screening and detection; VA facility-level data will be used to assess structural characteristics associated with screening. Aim 2 will examine three potential EA suspicion AMs and select the best performing via comparison to a multi-component reference standard consisting of: a) simplified rule-based NLP of progress note content, and b...