In the first phase of this study, a data-base containing clinical and laboratory findings of 704 patients with systemic lupus erythematosus (SLE), originating from 29 centres and 14 countries, was used to assess the validity of 4 common indices of disease activity, SLAM, BILAG, SLEDAI and SIS. The physician's judgement of activity was assumed as the unique reference criterion (gold standard). Computer programmes were developed to calculate automatically the 4 activity indices; this computation appeared to correspond with manual computations in a sample of 60 appropriately selected cases. All 4 indices were closely correlated with each other (r in the range of 0.716 to 0.872), and with the physician's score (r in the range of 0.620 to 0.719). In the second phase of the study the activity index developed in part I (ECLAM) was prospectively validated, and its performance compared to that of the other scales, both as a single state index and as a transition index (i.e., its ability to assess disease activity at a single point in time and to detect variations in consecutive readings). A computer-assisted clinical chart was prepared for this purpose. This chart allowed us to calculate automatically all the indices. Two consecutive observation times (time 0, and time 1 three months later) were included in the study protocol. Data on 75 patients from 19 centres were collected, and each patient was observed twice. All the computed indices were closely correlated, both at time 0 (r ranging from 0.725 to 0.884), and at time 1 (r ranging from 0.607 to 0.833).(ABSTRACT TRUNCATED AT 250 WORDS)