In health care, "person centeredness" is a valued (though nebulous) concept. In physical therapy, clinical interactions often strive to be person-centered, for example, by focusing on participation and valuing patient empowerment. However, the available evidence has mostly been constructed around populations (or study samples) rather than individuals. In this perspective, an alternative evidence framework is described, constructed around measurements in routine practice. Specifically, the authors propose developing "people-like-me" reference charts, generated with historical outcomes data, to provide real-Time information on an individual's status relative to similar people. The authors present an example of how this could work using their experience with people rehabilitating after total knee arthroplasty. They also describe several challenges that must be addressed to bring this innovation into practice. First, the most important outcome measures for stakeholders (eg, patients, clinicians) need to be identified and monitored longitudinally to ensure that "people-like-me" estimates are useful and support the goals of person-centered care. Statistical methods for selecting "people-like-me" need to be examined and refined. Finally, the "people-like-me" information needs to be packaged in such a way that it is accessible, intuitive, and helpful at the point of care. Ideally, the entire process should recognize from the outset that practice patterns evolve, so databases, statistical models, and decision tools should be dynamic by design. Ultimately, the authors propose this framework as a practical mechanism to advance person-centered decisions in physical therapy according to the ideals of evidence-based practice.