Risks of Medical MetaphorsTo the Editor In the A Piece of My Mind article, Ms Trogen 1 identified potential problems associated with the use of metaphor in the education of patients. Although we agree with her call for more effective communication with patients, we have reservations about the implications of extending the evidencebased paradigm to the use of metaphor.There is a conceptual error inherent in evidence-based practice when practitioners seek to apply population-level empirical data to make definitive claims about particular patients. 2 We suspect this shortcoming could be amplified in the application of population-level, "evidence-based" metaphors to individual patients. The use of effective and fitting metaphor in medicine requires attention to the particularities of patients' own stories, beliefs, experiences, and vernacular. As such, medical metaphors, as an extension of language, are inherently participatory and require a level of nuanced specification that other evidence-based practices do not. We have concerns that the application of evidence-based metaphors could therefore constrain the very adaptability that allows such speech to be both accessible and rhetorically compelling for patients.Metaphor is ubiquitous in human thought 3 and thrives on interpretation. Because of this, metaphors are particularly susceptible to conscription by institutions, which may result in limiting practitioner reasoning or patient autonomy and silencing novel or unusual language, particularly words and phrases that might challenge evidence-based norms. 4 If research determined that certain metaphors were more effective than others, it is possible that practitioner imaginations could be stultified by such evidence-based parameters, urging conformation to a normative metaphor that some practitioners may find unhelpful or even objectionable. The metaphors patients use could adapt to this evidence-based language, impeding their linguistic autonomy. Although medical metaphor is primarily understood as serving a communicative function, control of such language should be guarded against.We agree with Trogen's call for enhanced forms of patientphysician dialogue. As improved methods of communication in medicine are identified, the limits of such endeavors should be recognized.