Skip to main content

Table 1 The Area Under the ROC Curve for recruitment prediction test for each project. For each project the ROCAUC and 95% confidence intervals are given for the first and second cross validation datasets, followed by the average

From: Efficient identification of patients eligible for clinical studies using case-based reasoning on Scottish Health Research register (SHARE)

Project AcronymFold 1 (CI)Fold 2 (CI)Average (CI)
ALPHA0.947 (0.897–0.984)0.855 (0.777–0.928)0.901 (0.837–0.956)
ALLAY0.928 (0.903–0.955)0.854 (0.768–0.941)0.891 (0.836–0.948)
METFORMIN0.977 (0.970–0.989)0.980 (0.970–0.983)0.979 (0.970–0.986)
REFORM0.946 (0.896–0.980)0.957 (0.920–0.987)0.952 (0.908–0.984)
IMPOCT0.397 (NA)0.276 (NA)0.337 (NA)
TARDIS0.897 (0.844–0.946)0.887 (0.852–0.919)0.892 (0.848–0.933)
4P0.952 (0.909–0.993)0.744 (0.447–0.978)0.848 (0.678–0.986)
HF0.907 (0.851–0.957)0.927 (0.870–0.974)0.917 (0.861–0.966)
ImmunoStat0.632 (0.503–0.763)0.605 (0.443–0.756)0.619 (0.473–0.760)
  1. NA: insufficient data to calculate an accurate CI