A medical center launched a program to use patient genetic data to personalize medication dosing. Clinicians hoped that identifying variants affecting metabolism would reduce adverse reactions. Early results were encouraging: for a subset of patients, dose adjustments reduced side effects and shortened hospital stays. As the program expanded, administrators encountered a tension the initial proposal had not emphasized: the difference between clinical usefulness and data governance. Genetic data are durable; they can reveal information about family members and can be reinterpreted as science changes. Some patients consented to dosing decisions but were uncomfortable with long-term storage. Others worried that insurers might eventually obtain the data, despite current protections. The program’s ethics committee recommended separating functions. Genetic results needed for immediate dosing would be stored in the medical record with strict access controls, while raw sequence data would be stored separately with opt-in consent for future research. The committee also recommended plain-language explanations that distinguished “a result used to choose a dose today” from “data that could be reused for other purposes later.” The center’s final report concludes that precision medicine is not only a technical promise about better dosing. It is also a central challenge about trust and governance. Programs succeed, the report argues, when they integrate clinical benefits with transparent consent and storage choices that match what patients believe they are agreeing to.