PROJECT SUMMARY/ABSTRACT Older Americans rely on the emergency department (ED) for acute, unscheduled care. Unfortunately, many older adults experience poor outcomes after ED visits, suggesting that these encounters represent missed opportunities to identify high risk patients and intervene to improve the transition to outpatient care. In particular, significant falls among older adults are a serious and preventable problem. Unscheduled ED visits offer an opportunity to identify older adults at higher risk of falls than the general primary care population at a time when fall risk factors can be modified, and thus offer an ideal additional setting for fall screening beyond primary care. Such screening is advocated in the ED, but screening interventions often fail due to time constraints of providers in the emergency setting. As electronic health record (EHR) systems evolve, computerized decision support offers the potential to support fall screening with less provider burden. The objective of this proposal is to identify adults at high risk for future falls and improve their care both during their ED visit and after discharge. As a physician scientist, my goal is to lead an independent research program to improve transitions to outpatient care following ED visits for older adults. This 5-year proposal will advance these goals by providing the necessary support and training in implementation science as well as informatics-based interventions. I have a unique background in engineering, emergency medicine, and health services research and am well prepared to successfully complete the proposal. I will be aided by a team of expert mentors at an institution with substantial resources and an outstanding environment to conduct the research proposed and transition to an independent investigator with R01 support. The proposed aims are to: 1) Compare EHR-based data extraction to in-person screening of future outpatient fall risk, 2) Using data available in the EHR at the time of presentation, develop a predictive algorithm to risk- stratify ED patients for risk of significant falls in the next 6 months, and 3) Design and pilot a clinical decision support intervention to identify older adults at high risk of falls and improve their care both in the ED and after discharge. These aims will be accomplished by creating and analyzing a database linking the EHR and claims data, incorporating novel elements derived by natural language processing, by utilizing machine learning in addition to traditional statistical techniques, and by developing and piloting an intervention in one health system.