Methods. SLE patients were recruited into this multi-centre cross-sectional study. At every assessment, data were collected on disease activity and therapy. Logistic regression was used to model an increase in therapy, as an indicator of active disease, by the Classic BILAG score in eight systems. As both indicate inactivity, scores of D and E were set to 0 and used as the baseline in the fitted model. The coefficients from the fitted model were used to determine the numerical values for Grades A, B and C. Different scoring schemes were then compared using receiver operating characteristic (ROC) curves. Validation analysis was performed using assessments from a single centre.
Results. There were 1510 assessments from 369 SLE patients. The currently used coding scheme (A=9, B=3, C=1 and D/E=0) did not fit the data well. The regression model suggested three possible numerical scoring schemes: (i) A=11, B=6, C=1 and D/E=0; (ii) A=12, B-6, C-1 and D/E-0; and (iii) A-11, B-7, C-1 and D/E-0. These schemes produced comparable ROC curves. Based on this, A=12, B=6, C=1 and D/E=0 seemed a reasonable and practical choice. The validation analysis suggested that although the A 12, B=6, C=1 and D/E=0 coding is still reasonable, a scheme with slightly less weighting for B, such as A=12, B=5, C=1 and D/E=0, may be more appropriate.
Conclusions. A reasonable additive numerical scoring scheme based on treatment decision for the Classic BILAG index is A=12, B=5, C=1, D=0 and E=0.