Little high-quality research has been published regarding the comparative effectiveness of various treatments for prostate cancer,1 and in the face of this dearth of evidence, practice patterns for localized prostate cancer vary widely across individual practice sites regardless of cancer risk or other measurable patient factors.2 Jang and colleagues have completed a timely and important study investigating the associations among the types of physicians seen after diagnosis and ultimate primary treatment selection. Prior studies3 4 have examined issues of specialist bias toward their own treatment modality by means of surveys; this is the first to examine the actual impact of specialist visits on ultimate treatment selection.

The authors analyzed data from SEER-Medicare. These data are population based, though representative of only a subset of U.S. geographic regions. SEER data on prostate cancer cases are limited by an absence of PSA data and/or Gleason-grade data in about 40% of cases, and by an outdated classification of Gleason grade which obscures important differences in disease risk. This is an important limitation in that …