Data were drawn from the Adelphi RA Disease Specific Programme™ (DSP), a cross-sectional survey with elements of retrospective data collection administered to rheumatologists and their consulting patients in Germany, France, Italy, Spain, the UK), Japan, Canada, and the USA from June 2021 to February 2022. Orthopedists were included in Japan as they are heavily involved in the management and treatment of RA [23]. A complete description of the methodology has been previously described [24, 25], published [26], and validated [27], and has shown to be representative and consistent over time.

Rheumatologists completed a physician survey and then completed patient record forms (PRF) for the next 5–10 patients consulted in routine clinical practice. Patients for whom a PRF was completed were then invited to voluntarily complete a patient self-completion form. PRFs included information on patient demographics, RA diagnostic and treatment journey, disease activity, symptoms, management and treatment, concomitant conditions, and outcomes. Physicians were included if they were currently practicing as a rheumatologist (or an orthopedist in Japan), were actively involved in the management of patients with RA, making management decisions for ≥ 5 patients in a typical month, and were not recently qualified (≤ 1 year).

Patients were included if they had a physician confirmed diagnosis of RA, were ≥ 18 years of age, and were not participating in a clinical trial at the time of the survey.

The survey period continued until a quota of 5–10 patients was reached. The survey was fielded from June 2021 to January 2022 in Germany, France, Italy, Spain, the UK, and the USA; May 2021 to November 2021 in Canada; and September 2021 to January 2022 in Japan.

Physicians were also given the opportunity to complete a further 2–3 PRFs, reporting data for patients currently receiving UPA, to form an oversample. Given the recent approval of UPA in RA (between 2019 and 2021 in Europe [22], Japan [28], Canada [29], and the USA [21]), the oversample was carried out to ensure adequate sample size for analysis.

The population of interest for this study was patients with RA who were receiving a second line of AT at data collection for ≥ 6 months, having switched from an initial first line TNFi (n = 503); reasons for switching were reported. All patients were stratified by subsequent therapy of interest: TNFi-UPA (n = 261), TNFi-TNFi (n = 128), or TNFi-other MOA (n = 114). TNFi-TNFi excluded any patients who switched from originator to biosimilar for the same molecule. Patients with > 1 prior TNFi were also excluded.

Patient demographic and treatment history information collected included patient age, biological sex, Charlson Comorbidity Index (CCI), select cardiovascular (CV) risk factors (age ≥ 65 years, myocardial infarction, congestive heart failure, smoking status), time since diagnosis of RA, prior conventional synthetic disease-modifying antirheumatic drug (csDMARD) use, prior corticosteroid use, time to first TNFi, TNFi received as first-line treatment, duration of TNFi use, reasons for switching from TNFi, and duration of current treatment.

Clinical characteristics included physician-reported disease activity (with formal DAS28 scoring available for 29% of patients) categorized as follows: remission, < 2.6; LDA, 2.6 to < 3.2; moderate disease activity [MDA], ≥ 3.2–5.1; and high disease activity [HDA], > 5.1 at initiation of current treatment and the most recent follow-up (≥ 6 months from treatment switch). Physician-reported pain (none, mild, moderate, or severe) was also recorded at initiation of current treatment (baseline) and at the most recent follow-up. In addition, physicians reported fatigue (none, mild, moderate, or severe) and physician-reported adherence to medication (completely adherent or not completely adherent), which were recorded at the most recent follow-up. Physician-reported outcomes of disease activity (DAS28 scores), pain, fatigue, and medication adherence at most recent follow-up (≥ 6 months after treatment initiation) were compared for TNFi-UPA versus TNFi-TNFi and TNFi-UPA versus TNFi-other MOA.

Physicians were instructed to refer to the patients’ medical records when completing their PRF to ensure the most up-to-date medical information was included. DAS28 was collected using objective DAS28 assessments performed at the consultation, from medical records (DAS28 assessments performed within 4 weeks prior), or when neither of these was available, from the physician’s own clinical judgement. Within the analysis sample of n = 503 patients, 29.0% (n = 146) had a formal DAS28 score reported by the completing physician at that consultation or within 4 weeks prior. As data based on the physicians’ own clinical judgment were included, clinical outcomes were stratified by categories of physician-reported DAS28 scores (remission, LDA, MDA, HDA) at the most recent follow-up to assess if disease activity corresponded with severity of RA, physician-reported level of pain and various symptoms experienced (e.g., tender joints, swollen joints, stiffness in joints, loss of movement, sleep disturbance) (Supplemental Table S1). In addition, to ensure that physician clinical judgement aligned closely and was reflective of objective DAS28 assessment in clinical practice, Spearman’s correlation analysis was used to assess the correlation between physician-reported DAS28 and objective DAS28 scores. The results of this analysis indicated a strong correlation (Supplemental Table S2).

Descriptive summary statistics, including mean and standard deviation, were calculated for continuous variables. Frequency counts and percentages were calculated for categorical variables. For any missing values, patients were removed from all analyses in which that variable was used. There was no imputation of missing data. The number of patients with available data for each variable can be found in Table 1. Analyses were performed using UNICOM Intelligence Reporter, version 7.5.1 (UNICOM Systems, Inc. 2021. UNICOM Intelligence, Version 7.5.1. Mission Hills, CA: UNICOM Systems, Inc.) and Stata 17 (StataCorp, 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC).

The frequency and percentage of patients who achieved a change in physician-reported disease activity based on DAS28 (remission, < 2.6; LDA, 2.6 to < 3.2; MDA, ≥ 3.2–5.1; and HDA, > 5.1) from initiation of AT after TNFi to the follow-up visit for each treatment group were calculated.

Patient demographics and clinical characteristics were balanced using inverse-probability-weighted regression adjustment (IPWRA). IPWRA estimators use weighted regression coefficients to compute the averages of the outcome, which are converted to percentages (×100) for binary outcomes. The covariates balanced within the IPWRA were patient age, biological sex, CCI score, patients’ DAS28 at initiation of their current treatment as reported by their physician for the weighting and regression adjustment stage, and additionally current AT duration was used for the regression adjustment stage. Only patients who had data available for all variables were included in the IPWRA. Of the study population of 503, 493 (98%) were included in the adjusted analysis utilizing IPWRA; 10 patients were excluded because of missing data for DAS28 and pain at initiation of current advanced therapy treatment. Standard mean differences (SMDs) were reported for covariates before and after weighting.

Outcomes were modeled using IPWRA and reported as predicted percentages along with p values for each treatment group. IPWRA was conducted separately and compared between TNFi-UPA versus TNFi-TNFi and TNFi-UPA versus TNFi-other MOA.

Using a check box, patients provided informed consent for use of their anonymized and aggregated data for research and publication in scientific journals. Data were collected in such a way that patients and physicians could not be identified directly; all data were aggregated and de-identified before receipt. Data collection was undertaken in line with European Pharmaceutical Marketing Research Association guidelines, and therefore it does not require ethics committee approval. The study was performed in accordance with the Helsinki Declaration of 1964 and its later amendments.