I am trying to find the study that looked at surgery vs EM residents, chest tube placement and complications. I think it was out of Denver but I am not sure. Anyone remember the journal it was in?
The Kentucky paper is an exercise in how to lie with statistics.
"When the complication rate based on the operator who performed the TT was studied, the complication rate for tubes placed by physicians prior to arrival at our hospital was highest (38%). There was a complication rate of 13% for the 157 tubes placed by emergency physicians at our institution and a complication rate of 6% for the 355 tubes placed by surgeons Figure 2. The differences in the complication rates between prehospital tube placement vs emergency physician placement vs placement by surgeons were all significant (38% vs 13% vs 6%; P<.0001). Ninety-nine of the 355 chest tubes placed by surgeons were done in the emergency department, with a complication rate of 5% Figure 3, while tubes put in by surgeons in other areas of the hospital had a complication rate of 7%. To exclude multiple tube placements as a variable for differences in complication rates, we also examined the incidences of complications related to initial tube placement. When the initial chest tube was placed by a surgeon, the complication rate was 4.6% and when placed by an emergency medicine physician it was 11% (P=.0001)."
Notice a few things. First, their initial finding that "The differences in the complication rates between prehospital tube placement vs emergency physician placement vs placement by surgeons were all significant" compared all tubes placed by surgeons (including those in the OR) to the other physicians. They acknowledge that "tubes put in by surgeons in other areas of the hospital had a complication rate of 7%" but yet fail to direct compare tubes placed by surgeons in the ED for acute decompensation to those placed in the ED by EPs for the same. By always including "other" tubes in their data, we can't actually compare the complication rates. You have to wonder what that analysis would have shown (but I can guess! )