• lupus erythematosus, systemic
• outcome assessment, health care
• autoimmune diseases

We thank Chessa et al for their interest in our article on lupus Physician’s Global Assessment (PGA).1 Their correspondence discusses the measurement used to determine PGA scoring reliability and suggests that the intraclass correlation coefficient (ICC) used, ICC2,k, with a two-way random effect model based on mean scorings may have increased the reliability estimates reported, compared with a two-way random effect models based on individual data points (ICC2,1).2 Indeed correlation coefficients were approximately 39% lower for the pre-lab PGA and 27% lower for the post-labPGA when computed with the ICC2,1 compared with the ICC2,k. We used the ICC2,k to assess inter-rater reliability as our study was designed to evaluate and compare mean PGA scores before and after receipt of laboratory values from multiple (k=50) raters.3 Additionally, mean-based models (ICC2,k) are conventionally used to report ICC. Which of these models is more mathematically or conceptually correct remains debated. Importantly, both models showed increased agreement when laboratory values are available (post-lab PGA) compared with the pre-lab PGA. We wholeheartedly agree that further standardisation and precise guidelines for its scoring will likely improve the performance of the PGA when used by lupus experts and practising rheumatologists.