Background/Purpose: A physician global estimate of patient status (DOCGL) on a 0-10 visual analog scales (VAS) is as effective as any rheumatoid arthritis (RA) Core Data Set measure to quantify inflammatory activity in patients selected for clinical trials¹.  However, in routine care, DOCGL may be affected by joint damage and/or distress (e.g., fibromyalgia, depression, etc.), in addition to inflammation. Different physicians may consider these findings variably in formulating DOCGL. One approach to document the contribution of inflammation, damage, and distress to DOCGL is for the physician to complete 3 additional 0-10 VAS. We analyze inter-rater reliability between two rheumatologists for DOCGL and 3 VAS subscales for inflammation, damage, and distress.

Methods: At one academic site, 4 VAS for overall DOCGL, and levels of inflammation [or reversible findings] (DOCINF), organ damage [or irreversible findings] (DOCDAM), and distress [or symptoms not explained by inflammation or damage such as fibromyalgia, depression] (DOCSTR), are scored in patients with all rheumatic diagnoses in routine care. At a weekly fellows’ clinic, the four 0-10 VAS estimates scores were recorded independently by a senior rheumatologist and a rheumatology fellow at the same visit.  The estimates of the 2 observers for each subscale were analyzed for mean levels, differences, intraclass correlation coefficients (ICCs), and possible discordance – defined as a difference of 2 units/10 between the 2 observers.

Results: Estimates of 2 rheumatologists were analyzed in 107 patients with different rheumatic diseases. Mean differences ranged from 0.01 for DOCGL to 1.02 for DOCSTR, significant only for DOCSTR (Table).  The ICC ranged from 0.63 (DOCGL) to 0.80 (DOCDAM), a range higher than for joint counts². Concordance of estimates (within 2/10 units) ranged from 63% (DOCSTR) to 85% (DOCINF).

Conclusion: Good agreement between two observers with different levels of clinical experience was seen for 4 physician 0-10 VAS estimates for overall global assessment, inflammation, damage, and distress. These data extend evidence for the value of physician VAS, as DOCGL distinguishes active from control treatments in rheumatoid arthritis (RA) clinical trials more efficiently than joint counts or laboratory tests¹. Mean scores for damage and distress were higher than for inflammation, indicating the possible value for doctors, patients, and payers of quantitating damage, in addition to inflammation, to document patient status and support clinical decisions in routine care.

References: 1. Clin Exp Rheumatol. 2014;32 Suppl 85(5):47-54.  2.Ann Rheum Dis. 2009;68(6):972-5. 2. Bull NYU Hosp Jt Dis. 2008;66(3):216-23