We appreciate the comments of Dr Ghosh and agree that following evidence-based guidelines is a complex and often difficult task. In particular, we endorse Dr Ghosh's comments on the need to address both the expectations of patients and the way in which physicians present information on the risks of treatments.
Our article provides the first description of national antibiotic prescribing trends for uncomplicated urinary tract infections in women.1 We found that the first-line recommended antibiotic, trimethoprim-sulfamethoxazole, declined in use over the past decade while less-recommended and more expensive agents, fluoroquinolones and nitrofurantoin, increased dramatically in use during the same period. As suggested by Dr Ghosh, a variety of physician, patient, and health care system–level factors may explain these prescribing patterns. We uncovered differences in antibiotic prescribing between various specialists, suggesting that subspecialty culture plays a role. Similar differences among specialists have been observed previously.2 Further exploration of the reasons for these specialty differences, both qualitative and quantitative, may be a critical next step in understanding antibiotic prescribing patterns. One unique physician-level rationale for prescribing a second- or third-line choice may be a concern for emerging antimicrobial resistance, an issue that exists outside the traditional patient-physician relationship. Ideally, antibiotic recommendations for a community should be based on objective uropathogen resistance trends. Unfortunately, public reports of uropathogen resistance do not appear to be readily available to the clinician, and, when published, are specific to a particular region.3