Low interest in geriatrics training.

out that we do not know if physicians had asked patients to return after prescribing an antibiotic. There are certainly limitations to the use of medical billing data: (1) we could not measure the severity of a patient’s illness, (2) we were unable to capture any information about clinician recommendations for return visits, and (3) we did not collect ICD-9 codes for adverse events, which may be another plausible explanation for a patient returning for follow-up. However, even given these limitations, we believe several aspects of our study inclusion criteria helped address this issue, particularly for patients seen on the initial visit for a URI (ICD-9 465 and 460). To be classified as a URI for the purposes of our analysis, we excluded cases that were also diagnosed with any other respiratory illness at the time of the visit (such as acute otitis media, sinusitis, or pharyngitis), making it unlikely that they had a severe or complicated illness. We found that patients with acute otitis media, pharyngitis, and URI were more likely to return if they had received antibiotics on the initial visit. Although patients may be asked to return to see if their ear infection has resolved, it is not common practice in the United States to reassess patients with pharyngitis or URI. Thus, we believe that our finding suggesting that receipt of an antibiotic may actually increase return visits is plausible and deserves more study as part of evaluation of programs attempting to limit inappropriate use of antibiotics. John Li, MPH Department of Epidemiology University of Minnesota