A 2002 survey of 468 Canadian orthopedic surgeons found that the "two principal reasons regional anesthesia is not favored" are "delays in operating rooms" and "unpredictable success." We reanalyzed the data from the study to evaluate whether these concerns were the best predictors of an individual surgeon's willingness to use peripheral nerve blocks for their patients. Of the five procedures included in the survey, three had relevant questions for our reanalysis of the results: arthroscopic shoulder surgery, arthroscopic anterior cruciate ligament reconstruction, and total knee replacement. A surgeon's preference for peripheral nerve block for him or herself strongly predicted his or her anesthetic preference for patients (all P < 0.001). Concordance rates were 89% for arthroscopic shoulder surgery, 87% for anterior cruciate ligament reconstruction, and 93% for total knee replacement. There was almost no incremental predictive value for the surgeon's preference for patients from the surgeon's perception of the times to perform a block (P > or = 0.27) or perception of block success rate (P > or = 0.30). There was also almost no direct predictive value for the surgeon's preference for patients from the surgeon's perception of the times to perform a block (Kendall's tau < or = 0.04, P > or = 0.28) or perception of block success rate (Kendall's tau < or = 0.02, P > or = 0.24). An economically important percentage of surgeons (37%, 95% confidence interval: 32%-41%) would choose a peripheral nerve block for their own surgery for some, but not all, of the procedures (i.e., for 1 or 2 versus 0 or 3). A surgeon's preference for peripheral nerve blocks for his or her own surgery predicted a surgeon's preference for his or her patients. Perceptions of delays and success rate did not add sufficient incremental information to the surgeon's preferences to be of economic importance. These results are important to better forecast the net economic impact on an anesthesia group of a regional block team.