Results
Demographic characteristics of the CPG authors are described in Table 1. Among the 27 CPG authors of the JCR CPG for rheumatoid arthritis, 19 (70.4%) were male, 22 (81.5%) were affiliated with universities or university teaching hospitals, 9 (33.3%) were university professors, and 18 (66.7%) specialized in rheumatology and immunology. Only one author represented a rheumatoid arthritis patient organization. The median h-index and the number of academic publications were 21 (interquartile range [IQR]: 10.5–31.0) and 95 (IQR: 35–161), respectively.
Table 2 shows the personal payments from pharmaceutical companies to the CPG authors. All 27 authors received at least one personal payment from pharmaceutical companies between 2016 and 2020. A total of 3,951 payments, amounting to $3,683,048, were made to the CPG authors by 52 pharmaceutical companies over the five years. Speaking compensation accounted for 83.9% of the personal payments ($3,089,971), while consulting payments made up 10.3% ($380,217) of the total. The median and mean payments per author were $101,624 (IQR: $16,221–$184,160) and $136,409 (standard deviation [SD]: $156,736), respectively. The median and mean annual personal payments per author were $21,085 (IQR: $5,402–$37,968) and $29,000 (SD: $32,035), respectively. Gini index for per-author personal payments was 0.55. One author received $686,024 in personal payments over the five years. More than 77.8% (21 authors), 66.7% (18 authors), and 51.9% (14 authors) received more than $10,000, $50,000, and $100,000 in total payments over the five years, respectively. Although the CPG stated that all authors had self-declared their financial COI status to the JCR and that a COI management committee of the JCR had confirmed all authors with financial COIs as eligible for CPG development, this self-declared COI information was not disclosed in the main text of the CPG nor on the JCR’s website. Therefore, we could not evaluate the accuracy of the authors’ self-declared COI status.
Regarding the CPG recommendations, a total of 55 unique recommendations were listed. Of these, 15 (27.3%) were strong recommendations, and 40 (72.7%) were conditional recommendations. More than 81.8% (45 out of 55 recommendations) were supported by low or very low-quality evidence. Of the 15 strong recommendations, 66.7% (10 out of 15) were supported by low or very low-quality evidence, while 87.5% of conditional recommendations were based on low or very low-quality evidence.