| Dulchavsky SA et al, 2001, USA | 382 stable surgical patients (95% post-traumatic) | Prospective diagnostic study | USS v CXR diagnosis | Sensitivity 95%(89–95)Specificity 100% | Only stable patients recruited. CXR used as gold standard |
| | | | (99–100) | |
| Rowan KR et al, 2002, Canada | 27 patients sustaining blunt chest trauma who had CT scans | Prospective blinded diagnostic study | USS v CT diagnosis | Sensitivity 100%(82.6–100)Specificity 94% | May have selection bias for large pneumothoraces Small numbers |
| | | | (82–94) | |
| Knudston JL et al, 2004, USA | 328 consecutive trauma patients | Prospective diagnostic study | USS v CXR diagnosis | Sensitivity 92.3%(74.4–97.9)Specificity 99.7% | CXR used as gold standard. Not clearly blinded |
| | | | (98.9–99.9) | |
| Kirkpatrick AW et al, 2004, USA | 225 trauma patients | Prospective diagnostic study | USS v CT diagnosis or escape of air on thoracostomy | Sensitivity 58.9%(45.0–71.9)Specificity 99.1% | Unclear if CT radiologists blinded to USS |
| | | | (97.6–99.8) | |