Classification criteria of joint activity using joint index vector for patients with rheumatoid arthritis: An evaluation and verification

Abstract

Objectives

We developed an activity classification of the joint index vector (JIV) for rheumatoid arthritis (RA) using monitoring data of RA cases at our institute. We verified its validity using an external big dataset (BD).

Methods

JIV is a novel joint involvement evaluation method that presents three-axis coordinates. We have set JIV classification criteria to determine a cut-off index (COI) of the combined vector on the x-y axis (Vxy). The z-axis (Vz) in the JIV was determined by Receiver Operating Characteristic analysis (ROC) in referring to the Clinical Disease Activity Index (CDAI) disease activity threshold and Health Assessment Questionnaire Disability Index (HAQ-DI) remission criteria. The criteria of JIV were evaluated in relation to indicators such as the CDAI and HAQ-DI. After determining the criteria, the validity was verified by referring to the simplified disease activity index (SDAI) classification and the HAQ score in BD.

Results

A total of 617 patients were studied. These were defined as 0.1>Vxy as remission (REM), 0.45>Vxy≥0.1 and 0.125≥Vz as low joint activity (LJA), 1.0>Vxy≥0.45 or 0.45>Vxy and Vz>0.125 as moderate joint activity (MJA), and Vxy≥1.0 was defined as high joint activity (HJA). The external big dataset consists of 11,013 RA patients. Vxy and the SDAI score correlated significantly (p < 0.0001). Mean values of SDAI and HAQ-DI increase stepwise as the criteria upgrade. It has been suggested that JIV may be able to pick up patients at risk of being missed by SDAI when large joints are involved.

Conclusions

JIV has been assessed and verified for its appropriateness, unbiased evaluation of the joints, and advantages in covering an unmet need in SDAI.