Patient-reported outcomes (PROs) reflect the patient’s perspective and are used in rheumatoid arthritis (RA) routine clinical practice. Patient global assessment (PGA) is one of the most widely used PROs in RA practice and research and is included in several composite scores such as the 28-joint Disease Activity Score (DAS28). PGA is often assessed by a single question with a 0–10 or 0–100 response. The content can vary and relates either to global health (e.g., how is your health overall) or to disease activity (e.g., how active is your arthritis). The wordings used as anchors, i.e., for the score of 0, 10, or 100 according to the scale used, and the timing (i.e., this day or this week) also vary. The different possible ways of measuring PGA translate into variations in its interpretation and reporting and may impact on measures of disease activity and consequently achievement of treat-to-target goals. Furthermore, although PGA is associated with objective measures of disease activity, it is also associated with other aspects of health, such as psychological distress or comorbidities, which leads to situations of discordance between objective RA assessments and PGA. Focusing on the role of PGA, its use and interpretation in RA, this review explores its validity and correlations with other disease measures and its overall value for research and routine clinical practice.

(249 words)Objective: The integration of the patient in therapeutic decision-making is important in the management of rheumatoid arthritis (RA); but the patient opinion regarding disease status may differ from the physician's opinion. The aim of this study was to assess in the published literature the frequency and drivers of patient-physician discordance in global assessment in RA.Method: Systematic literature review by 2 investigators of all papers published up to January 2015 in Medline or EMBASE, reporting discordance in RA. Discordance was defined based on the absolute difference of patient global and physician global assessments (PGA/PhGA) on 0-10cm scales. The frequency of discordance and its predictors were collected in each study. Frequencies of discordance were pooled by metaanalysis using random effect. Results:In all, 12 studies were selected (i.e.,11,879 patients): weighted mean age 55.1±13.9 years, weighted mean disease duration 10.4±9.3 years, 80.7% were women. The value of the difference |PGA-PhGA| defining discordance varied between ≥0.5cm (N=2 studies) to ≥3cm (N=5 studies); the weighted mean value was 2.7cm. The pooled percentage of patients with discordance was 43% (95% confidence interval 36%-51%, range: 25%-76%). PGA was usually higher than PhGA. The drivers of PGA were pain and functional incapacity, whereas drivers of PhGA were joint counts and acute phase reactants. Conclusion:Discordance in global assessment was most frequently defined as a difference of 3 points or more; even with such a stringent definition, up to half the patients were found to be discordant. The long-term consequences of this discordance remain to be determined.