Skip to main content
. 2013 Aug 21;8(8):e71154. doi: 10.1371/journal.pone.0071154

Table 2. The proportion correctly identified by severity of abnormality as well as the sensitivity, specificity and area under the ROC curve (AUC) for each study design in trials 1 and 2 by classification difficulty.

Trial 1 Trial 2
Proportion correctly identified Proportion correctly identified
0.03c 0.05c 0.03c_500_90% 0.03c_5000_99% 0.03c 0.05c 0.03c_500_90% 0.03c_5000_99%
Mildly abnormal (N = 1200) 57% 64% 55% 72% 67% 69% 67% 79%
Normal (N = 600) 77% 75% 87% 64% 87% 86% 89% 52%
Severely abnormal (N = 200) 96% 92% 90% 99% 99% 96% 98% 99%
Specificity Specificity
0.03c 0.05c 0.03c_500_90% 0.03c_5000_99% 0.03c 0.05c 0.03c_500_90% 0.03c_5000_99%
Normal-Mildly abnormal (Difficult classification) 74% 68% 85% 64% 87% 86% 89% 52%
Normal- Severely abnormal (Easy classification) 74% 68% 85% 64% 87% 86% 89% 52%
Normal-Abnormal (All) 74% 68% 85% 64% 87% 86% 89% 52%
Sensitivity Sensitivity
0.03c 0.05c 0.03c_500_90% 0.03c_5000_99% 0.03c 0.05c 0.03c_500_90% 0.03c_5000_99%
Normal-Mildly abnormal (Difficult classification) 61% 70% 61% 72% 67% 69% 67% 79%
Normal- Severely abnormal (Easy classification) 98% 99% 98% 99% 99% 96% 98% 98%
Normal-Abnormal (All) 66% 74% 66% 76% 72% 73% 72% 82%
Area under the ROC curve (AUC) Area under the ROC curve (AUC)
0.03c 0.05c 0.03c_500_90% 0.03c_5000_99% 0.03c 0.05c 0.03c_500_90% 0.03c_5000_99%
AUC (95%CI): Normal-Mildly abnormal (Difficult classification) 0.678(0.656–0.700 ) 0.692(0.669–0.715 ) 0.731(0.711–0.751 ) 0.681(0.658–0.704 ) 0.771(0.752–0.790) 0.777(0.758–0.796) 0.784(0.766–0.802) 0.656(0.634–0.680)
AUC (95%CI): Normal-Severely abnormal (Easy classification) 0.871(0.850–0.889) 0.833(0.813–0.854 ) 0.915(0.895–0.930) 0.819(0.799–0.839 ) 0.929(0.913–0.944) 0.91(0.891–0.929) 0.938(0.922–0.953) 0.754(0.732–0.776)
AUC (95%CI): Normal-Abnormal (All) 0.704(0.683–0.724) 0.712(0.692–0.732) 0.757(0.738–0.776) 0.701(0.680–0.721) 0.794(0.776–0.811) 0.796(0.778–0.814) 0.806(0.789–0.823) 0.671(0.648–0.693)

(0.03c = study design 1; 0.05c = study design 2; 0.03c_500_90% = study design 3; 0.03c_5000_99% = study design 4).