ABSTRACT Despite assumptions that inter-hospital transfer (IHT, the transfer of patients between acute care hospitals) is done to provide patients with necessary specialized care, IHT also exposes vulnerable hospitalized patients to the risks of discontinuity of care. Our recent research suggests that a subset of transferred patients may experience harm during IHT without clear benefit; however, we currently lack data on how to identify in which patients harm outweighs benefit of transfer, i.e., experience inappropriate IHT. The long-term objective of this research is to rigorously define and evaluate the incidence and patient safety impact of potentially inappropriate IHT among hospitalized medical patients, identify patients at risk for inappropriate transfer, and develop an intervention toolkit to reduce potentially inappropriate IHT. To achieve this, we will work with diverse stakeholders involved in IHT, including patients/families, accepting and admitting clinicians, transferring clinicians, and hospital leadership to define potentially inappropriate IHT. We will then determine the incidence and patient safety impact of potentially inappropriate IHT via standardized adjudication of 1800 hospitalized medical patients from 18 US hospitals that participate in a national research collaborative (HOMERuN) and contribute data to a benchmarking and purchasing organization (Vizient). Adjudication tools will be based on those used in similar prior research and will be informed by stakeholder definitions of potentially inappropriate IHT. We will incorporate rigorous adjudicator training and continuous review to ensure reliability across sites, adjudicators, and time. We will then use standard modelling techniques to retrospectively characterize a population(s) of potentially inappropriate patient transfers based on several patient, transfer process, and system-level factors. This will be followed by advanced machine-learning methods to prospectively identify patients at risk for potentially inappropriate IHT and validation of the model's performance. Finally, we will utilize key results from the above analyses and stakeholder input to create a prototype intervention, refine the prototype based on feedback from participating sites using a mixed methods approach, and develop a toolkit of best practices to prevent potentially inappropriate IHT (e.g., by replacing it with a safer alternative) for future dissemination. This research will establish a foundation from which healthcare systems can achieve excellence in providing patients with the right care, in the right setting.