Understanding patient and physician global assessments in rheumatoid arthritis: A meta-epidemiological study of core outcome set performance.

OBJECTIVE

To assess agreement between physician and patient global assessment improvements in randomised controlled trials (RCT) of targeted therapies for rheumatoid arthritis and to identify which core outcome domains best explain changes in each assessment and their discrepancies.

METHODS

We conducted a systematic review and meta-epidemiological analysis of RCTs comparing targeted therapies with placebo. Standardised mean differences (SMDs) were calculated for all eight RA-COS domains: Physician global assessment, Patient global assessment, Disability, Tender joint count (TJC), Swollen joint count (SJC), Pain, Acute phase reactants, and Fatigue. Agreement between physician and patient assessments was assessed as ΔSMD Global (i.e., SMD Physician global assessment - SMD Patient global assessment). Multivariate meta-analyses were performed to calculate SMD for the remaining domains. Multiple imputation followed by meta-regression analyses identified domain-level predictors of global assessment changes.

RESULTS

We identified 73 double-blind RCTs (165 randomised comparisons) involving 33,956 RA patients. Physician assessments demonstrated slightly greater treatment effects (SMD -0.60; 95% CI -0.65 to -0.55) compared with patient assessments (SMD -0.50; 95% CI -0.54 to -0.46): ΔSMDGlobal = -0.09 (95% CI -0.12 to -0.05). Univariable meta-regression identified disability, TJC and SJC as the strongest associations of both assessments, yet SJC was the only domain significantly associated with the contrast between these (ΔSMDGlobal). Pain was independently associated only with patient global assessment.

CONCLUSION

Targeted therapies show larger improvements in physician-assessed outcomes than on patient-reported outcomes, largely driven by SJC. Pain remains a distinct determinant of patient assessments, underscoring the need to better integrate patient-reported outcomes into RA trial design.

REGISTRATION: The protocol was registered on PROSPERO (2025-02-14).

REGISTRATION NUMBER: CRD420250652342.