Twelve studies were included in the review.
Heterogeneity between the studies was found to result primarily from whether or not the reference standard independently reported radiographs and from the diagnostic threshold employed.
A funnel plot demonstrated no evidence of publication bias in the review.
Accuracy of radiographer plain radiograph reporting.
The summary sensitivity of radiographer reporting was 92.6% (95% CI: 92.0, 93.2), with a specificity of 97.7% (95% CI: 97.5, 97.9). ROC curves indicated that radiographer reporting accuracy in clinical practice was comparable to that of the reference standard.
Heterogeneity was statistically significant (for both sensitivity and specificity; P<0.001). The pooled DOR using the DerSimonian and Laird random-effects model was 540.9 (95% CI: 303.4, 965.3), with a high level of heterogeneity (P<0.00001). An asymmetrical ROC curve fitted using the Littenburg-Moses model gave some indication that the DOR was dependent on the diagnostic threshold.
Accuracy of selectively trained radiographers and radiologists of varying seniority.
Based on two studies, there was no evidence of a difference in the accuracy of reporting of AED radiographs for all body areas compared with a reference standard.
Accuracy of selectively trained radiographers reporting for different types of patient and body areas.
There was no evidence of a difference between how accurately trained radiographers report AED radiographs when reporting radiographs not solely from an AED. There was no difference in the summary estimates of sensitivity for different body areas; however, specificity was higher for the axial than the appendicular skeleton.
The effect of training on the accuracy of radiographer reporting.
There was no significant difference in estimates of sensitivity for AED radiographs between radiographers who had received training and those who had not. However, there was a difference in specificity for all body areas in favour of those who had received training.