Eric Weinberg, Michael Tunik, James Tsung; International Journal of the Care of the Injured; 2010; 41; 862-868.

The authors are ED types who used video and simulated fractures to teach a group of ten ED physicians to use the ultrasound to identify fractures. In one hour! The novice sonologists scattered to two different EDs and scanned 212 patients with possible fractures.
The results: the novices had an overall sensitivity of 73% and specificity of 92% of making an accurate diagnosis of fracture (all patients had X-rays or CT scans read by attending radiologist as the gold standard). The novices did better with long bones and skulls than with articular fractures.

Our take home points:

  1. The study used radiologist interpreted X-rays and CT scans as the gold standard, but that standard is neither 100% sensitive nor 100% specific. A randomized controlled trial of treatment outcomes based on ultrasound versus x-ray supported clinical judgment would be interesting.
  2. Sensitivity should improve with practice (supported by study outcomes). Too many false negatives are not good when trying not to miss a fracture.
  3. Novices should not use ultrasound for making the call of Salter-Harris/periarticular fractures.
  4. Ultrasound is a promising, mobile tool, with a short learning curve. It is particularly well-suited to limited resource environments for the assessment of suspected non-articular long bone and skull fractures.