Agreement Between Physician Evaluation and the Composite Response Index in Diffuse Cutaneous Systemic Sclerosis

Objective Diffuse cutaneous systemic sclerosis (SSc) is a highly heterogeneous disease. A provisionally approved Composite Response Index in diffuse cutaneous SSc (CRISS) was developed as a 1-year outcome measure for clinical trials. Our goal was to further validate the CRISS by examining agreement between CRISS definitions for improved/non-improved with physicians' evaluation of disease. Methods Patient profiles from a large observational cohort were created for 50 random diffuse cutaneous SSc patients of <5 years disease duration with improved CRISS scores after 1 year and 50 with non-improved CRISS scores. Profiles described disease features used during the initial CRISS development at baseline and at 1 year. Each profile was independently rated by 3 expert physicians. Majority opinion determined whether a patient was improved or not improved, and kappa agreement with the CRISS cutoff of 0.6 was calculated. Results Patients had mean +/- SD disease duration of 2.2 +/- 1.3 years. There was substantial agreement between the physician majority opinion about each case and the CRISS (kappa = 0.76 [95% confidence interval (95% CI) 0.64-0.88]). The agreement between each individual physician opinion and the CRISS was also substantial (kappa = 0.70 [95% CI 0.62-0.78]). All CRISS non-improvers were also rated as non-improved by physician majority; however, 12 CRISS improvers were rated as non-improved by physicians. Conclusion There was substantial agreement between the dichotomous CRISS rating and physician assessment of diffuse cutaneous SSc patients after 1 year. This supports the use of a CRISS cutoff at 0.6 for improvement versus non-improvement, although the CRISS tended to rate more patients as improved than did physicians.

Automated summary

The study aimed to validate the Composite Response Index in diffuse cutaneous systemic sclerosis (CRISS) by examining the agreement between CRISS definitions and physicians' evaluation of disease. The results showed substantial agreement between CRISS rating and physician assessment of patients, supporting the use of CRISS cutoff at 0.6 for improvement versus non-improvement.