The patient's voice is increasingly present in various contexts. I focus on two clinical contexts, psychoanalysis and narrative medicine, and consider the clinical impact of speaking and being listened to in the epistemic context of patient-based research. I first consider the relation between epistemic and clinical values within clinical contexts and show that words are not healing in virtue of their epistemic value, neither in psychoanalysis, nor in narrative medicine. I then underline that, in different ways in psychoanalysis and narrative medicine, the healing power of words is tied to the recognition of the speaker as the owner of his own voice. Listening singularly to the subject, however, conflicts with academic research, as the latter imposes to make the patient's experience fit in the standardized format appropriate for the generation of data, and requires the anonymity of the patient participating to research. Over and above these tensions between clinical and epistemic contexts, the use of narratives both in medicine and patient-based research have in common to rely on an ethic of mastery, aiming at domesticating meaningless illness with meaningful words. Even though such process may be psychologically gratifying, it relies on an illusion masking the patient's vulnerability, inherent to his sufferance, and pressures him to neglect aspects of his experiences which cannot be narrated: in narrative contexts, meaninglessness remains unheard. These difficulties are inherent to narrative approaches,both in clinical and epistemic contexts, but may not occur in psychoanalysis. I conclude by underlining that, in healing or harming ways, addressing oneself to another is a transformative experience insofar as it involves speaking and being listened to as a singular subject. This is the case not only in clinical but also in epistemic contexts. Therefore, patient-based research is not and never clinically inert: the patient-based researcher should always be clinically responsible.