Overprescription of opioid pain relievers is a known contributor to the growing opioid epidemic. Identification of medical providers that engage in overprescription has proven challenging. We examine the utility of physician rating websites (PRWs) as potential sources of data that may help identify overprescribing practices. We leverage text mining techniques to identify linguistic cues that are associated with known cases of overprescription. We find that patients flag potentially problematic medical providers in their reviews and suggest that intervention by authorities is warranted. Our study contributes to the growing body of literature on medical infoveillance by identifying patients' appeals to regulatory authorities as an important type of social signal for regulatory monitoring.