An attempt and validation for classification of disease activity of rheumatoid arthritis using the joint index vector: part2

Abstract

［Objectives］We evaluated the validity of the activity classification of Joint Index Vector（JIV）using a big data that already published.

［Methods］Classification criteria were created using the method used to create cJIV based on the processed NinJa 2015 data used to evaluate JIV for comparison with the existing classification（NcJIV）. The indicators on which the criteria are based are the disease activity indicator of the Simplified Disease Activity Index（SDAI）and the remission criteria of the Health Assessment Questionnaire Disability Index（HAQ-DI）. We compared NcJIV and cJIV, and also compared indicators such as SDAI and HAQ-DI.

［Results］We analyzed 11013 of NcJIV and 617 of cJIV datasets. There were defined as 0.05>Vxy as remission（REM）, 0.25>Vxy≥0.05 as low joint activity（LJA）, 0.7≥Vxy≥0.25 and 0.225≥Vz as low-middle joint activity（LMJA）, 0.7≥Vxy≥0.25 and Vz>0.225 as middle joint activity（MJA）, and Vxy>0.7 was defined as high joint activity（HJA）. The concordance rate（susceptibility）of cJIV to NcJIV was 100% for REM, 79.5% for LJA, 80.3% for MJA, and 93.0% for HJA.

［Conclusions］cJIV is highly sensitive even in big data, suggesting it is a highly versatile classification standard.