Results
| No of patients (N=) | Sensitivity | Specificity | Positive predictive value | Negative predictive value | False-positive | False negative | Risk of bias | |
| % | % | % | % | N= | N= | |||
| (95% CI) | (95% CI) | (95% CI) | (95% CI) | |||||
| High percentage of fractures | ||||||||
| Parri35 | 55 | 100 (88 to 100) | 95 (73 to 100) | 97 (84 to 100) | 100 (79 to 100) | 1: Non calcified suture ipsilateral to trauma | 0 | Low risk of bias |
| Parri40 | 115 | 91 (82 to 96) | 85 (65 to 95) | 95 (88 to 98) | 74 (55 to 87) | 4: Not reported | 8: Fracture not underneath area that was imaged | At risk of bias |
| Low percentage of fractures | ||||||||
| Weinberg36 * | 21 | 100 (20 to 100) | 100 (79 to 100) | 100 (20 to 100) | 100 (79 to 100) | 0 | 0 | At risk of bias |
| Riera37 | 40 | 67 (24 to 94) | 97 (83 to 100) | 80 (30 to 99) | 94 (79 to 99) | 1: Not reported | 1: Not reported | At risk of bias |
| 1: No cooperation of patient | ||||||||
| Rabiner38 * | 69 | 88 (47 to 99) | 97 (88 to 99) | 78 (40 to 96) | 98 (90 to 100) | 1: Novice error | 1: Fracture adjacent to haematoma in the area that was not imaged | At risk of bias |
| 1: CT no fracture, possibly false negative as the sonographers seem to be convinced that it truly was a fracture | ||||||||
| Choi39 | 87 | 77 (46 to 94) | 100 (94 to 100) | 100 (66 to 100) | 96 (88 to 99) | 0 | 2: Difficult evaluation in area of orbital wall and skull base | At risk of bias |
| 1: Fracture not in imaged area, adjacent to haematoma | ||||||||
| Masaeli34 | 538 | 92 (83 to 97) | 96 (94 to 97) | 79 (69 to 87) | 99 (97 to 99) | 19: Not reported | 6: Not reported | Low risk of bias |
| Pooled data | ||||||||
| Overall pooled data | 925 | 91 (87 to 94) | 96 (94 to 97) | 88 (84 to 92) | 97 (95 to 98) | 27 | 20 | |
| Pooled data A | 170 | 93 (87 to 97) | 89 (76 to 96) | 96 (90 to 98) | 84 (70 to 92) | 5 | 8 | |
| Pooled data B | 755 | 89 (81 to 94) | 97 (95 to 98) | 81 (73 to 88) | 98 (97 to 99) | 22 | 12 | |
| Pooled data C | 720 | 92 (85 to 96) | 96 (95 to 98) | 85 (78 to 90) | 98 (97 to 99) | 21 | 11 | |
| Pooled data D | 835 | 91 (87 to 95) | 96 (94 to 97) | 88 (84 to 92) | 97 (95 to 98) | 25 | 19 | |
Pooled data A (studies with a high percentage in fractures): Parri35 and Parri40.
Pooled data B (studies with a low percentage in fractures): Weinberg36, Riera37, Rabiner38, Choi39 and Masaeli34.
Pooled data C (studies with a low percentage in fractures, excluded the studies that contained patients aged >18 years): Riera37, Choi39 and Masaeli34.
Pooled data D (all studies, excluded the studies that contained patients aged >18 years): Riera37, Parri35, Parri40, Choi39 and Masaeli34.
*Study that contained patients aged >18 years.