Performance of formal and informal prediction on a dichotomous split (OHS≥3) and across an ordinal OHS (defined on five levels: 0, 1, 2, 3 and ≥4).

NOTE: The ORC is a measure of discrimination for ordinal models ranging from 0.5 (no discrimination) to 1 (perfect discrimination). ABBREVIATIONS: Ordinal c-index (ORC); Oxford Handicap Scale (OHS) and the Six Simple Variables model (SSV). Note that all confidence intervals are 95% CIs. The ORC (an ordinal equivalent to the AUROCC) and the sensitivities/specificities CIs are calculated over 1000 bootstrap replicates within a single imputation of the ESS and the doctors’ sensitivity and specificity CIs are Zhou-Li intervals.Performance of formal and informal prediction on a dichotomous split (OHS≥3) and across an ordinal OHS (defined on five levels: 0, 1, 2, 3 and ≥4).