Background ESSDAI is a validated activity index in primary Sjögren's syndrome more and more used in clinical trials but also for finding biomarkers associated with activity of the disease. In this context, the presence of the biological domain in the ESSDAI score may induce circular reasoning and colinearity of data.
Objectives To develop the ClinESSDAI, derived from the ESSDAI score after deleting biological domain, and to validate if by comparison with the original ESSDAI.
Methods The 702 fictive vignettes created using data from 96 real cases of pSS for the study of development ESSDAI were used. As for the development of the original score, the assessment by the 39 experts who participated to the development study of disease activity using a scale ranging from 0-10 was used as the “gold standard” for weighting domains in the robust regression model. The explanatory variables included all domain of ESSDAI except the biological domain. The ClinESSDAI was compared to original ESSDAI to assess if it could be a surrogate of the original score by assessing its reproducibility using ESSDAI as gold standard. The validity of ClinESSDAI (including construct validity, reliability and sensitivity to change) was then assessed and compared to that of ESSDAI. Validation populations included the 96 real cases and the 395 patients of the EULAR cohort.
Results In the multivariate model, each of the 11 areas was significantly associated with disease activity. The weight of the domains was slightly different from the fields of ESSDAI (table). The ClinESSDAI was an excellent surrogate of the original ESSDAI, since the ICC between ESSDAI and ClinESSDAI were0.98 [0.97; 0.98] for fictive vignettes, 0.99 [0.98; 0.99] for the real case and0.90 [0.87; 0.92] in the EULAR cohort. Psychometric properties of the ClinESSDAI (construct validity, reliability and sensitivity to change) were very close to that of ESSDAI.
Conclusions Declination of the ESSDAI score without biological domain appears valid and very close to the original score. In this score, weights of some domains changed. This score will allow an evaluation of disease activity that is independent ofB-cell biomarkers, which will be useful in biological studies to avoid circular reasoning and colinearity of data. In addition, it may be helpful in clinical practice to assess disease activity when immunological tests have not been carried out.
Disclosure of Interest None declared