Outcomes Following Adalimumab Bio-originator to Biosimilar Switch—A Comparison Using Real-world Patient- and Physician-Reported Data in European Countries

Survey Design

Data were drawn from the Adelphi Rheumatoid Arthritis Disease-Specific Programme™ (DSP) conducted in 2020 in France, Germany, Italy, Spain, and the United Kingdom. DSPs are large, multinational, point-in-time surveys conducted in real-world clinical practice of physicians and their consulting patients. A complete description of the methods of the DSP has been previously published and validated [19,20,21].

Participants

A geographically representative sample of physicians was recruited to participate in the DSP. Physicians were eligible to participate in the survey if they were personally responsible for and actively involved in treatment decisions and management of patients with RA, and consulted at least three patients with RA in a typical month.

Patients were eligible for inclusion if they were over 18 years of age, had a physician-confirmed diagnosis of RA and visited the physician. Patients included in this analysis are a combination of a sub-set of these consecutively sampled patients, and a deliberately captured additional set of patients (the over-sample). These patient samples were chosen to assess the effects of switching from ADA originator to an ADA biosimilar. The consecutively sampled patients were the next ten prospectively consulting RA patients seen by their physician. Those who were either currently being prescribed ADA bio-originator (referred to as non-switchers) or currently being prescribed ADA biosimilar therapy, having switched from ADA bio-originator therapy (referred to as switchers), were included for analysis. Physicians selected which ADA biosimilar the patient was taking from a list of available brands at time of data collection. The over-sample included switchers only. Both consecutively sampled patients and the over-sample of switchers had to fulfil the inclusion criteria. Patients who had received advanced therapy prior to initiation of ADA bio-originator were removed from the sample to remove any confounding effects of previous advanced therapy use. Switcher patients were not requested to have a minimal time period since switch from ADA originator to an ADA biosimilar.

“Advanced therapy” was used to collectively refer to the following: targeted synthetic DMARDs (tsDMARD; e.g., tofacitinib, baricitinib, upadacitinib), biologic originator DMARD (boDMARD; e.g., abatacept, adalimumab, certolizumab-pegol, etanercept, golimumab, infliximab, rituximab, sarilumab, and tocilizumab), and biosimilar DMARD (bsDMARD; e.g., those for etanercept and infliximab).

Data Collection

Physicians (rheumatologists) completed a patient record form (PRF) for their next ten consecutive patients with RA who visited their clinic for routine care. The PRF contained questions on demographics, treatment history, and clinical outcomes such as satisfaction and adherence. Completion of the PRF was through consultation of existing patient clinical records and, consistent with decisions made in routine clinical practice, the judgement and diagnostic skills of the respondent physician. Thereafter, the same rheumatologists provided data for the over-sample by completing up to a further four PRFs on a prospective basis.

Disease activity, satisfaction, and patient adherence were collected as physician-reported outcomes. Disease severity was assessed by physicians’ subjective opinion as mild, moderate, or severe across key timepoints from initiation of current ADA treatment to the time of data collection. Patients were considered to have either improved, worsened, or remained stable based on physician-perceived severity (mild, moderate, severe) across these timepoints. Disease severity change was measured for two time frames: firstly, from initiation of first ADA treatment to time of data collection, secondly, from initiation of current ADA treatment to time of data collection.

Physicians invited the patients for whom PRF was completed to complete a patient-reported self-completion form (PSC). The PSC collected data on patient demographics, compliance, treatment satisfaction, and QoL. The PSC also asked about the impact of RA on patients’ lives, using the EuroQol- 5 Dimension (EQ-5D) [22] and the Work Productivity and Activity Impairment (WPAI) questionnaire [23]. Patients’ pain and fatigue were assessed using the Health Assessment Questionnaire (HAQ) [24] and the Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-F) [25, 26], respectively.

All physician-reported and patient-reported questions and scoring are detailed in Supplementary Table S1.

Ethical Approval

The survey did not require Ethics or Institutional Review Board review due to its non-interventional, observational nature of data collection. Informed consent from physicians and patients was required before their participation. The survey was conducted in accordance with the Western Institutional Review Board (protocol number 21-ADRW-104). Where patients provided data directly, they signed an informed consent form prior to participation in this study.

Data collection was conducted in accordance with national market research and privacy regulations, including European Pharmaceutical Market Research Association (EphMRA), and the US Department of Health and Human Services National Institutes of Health, Health Insurance Portability and Accountability Act (HIPAA).

All responses captured on the data collection forms were anonymized to preserve respondent confidentiality. Responses were anonymized before aggregated reporting, the identity of the physicians was blinded, and no patient identifiers were collected.

This study was performed in accordance with the Helsinki Declaration of 1964, and its later amendments.

Statistical Analysis

Data were summarized using descriptive analyses using Stata 17 [27]. Means and standard deviations (SD) were calculated for continuous variables, and frequency and percentages calculated for categorical variables. For descriptive data, numbers and percentages were shown in each category. Where propensity score matching was conducted, relevant scores (means or percentages) for each group are shown with an associated p value. Sample n numbers are not shown, as these values represent the scores of a matched patient population, not a true population.

Initially, bivariate comparisons of the two patient populations, the 186 non-switchers and the 284 switchers, were used to compare clinical and QoL outcomes. These analyses showed that the baseline characteristics were significantly different between the two populations. Key differences across the two populations included: a longer time since diagnosis of switchers (an additional 6.8 months), current ADA therapy duration was shorter for switchers (13.2 months less), fewer switchers with lower physician-reported disease activity based on DAS28 score estimates (remission [DAS28 < 2.6], low [DAS28 2.6–3.2], moderate [DAS28 3.3–5.1], and high disease activity [DAS28 > 5.1]) at diagnosis (p = 0.03), and more switchers with lower physician-reported disease activity currently (p = 0.202) compared with non-switchers.

To control for these differences, non-switchers and switchers were compared using propensity score matching (PSM), confounding for patient age, sex, body mass index, time since first ADA therapy (months), time since diagnosis (months), and the change in physician-perceived severity categories (mild, moderate, severe) over the duration of all ADA therapy (from initiation of ADA originator to the time of data collection). The data were matched in two ways: physician-reported patient data were matched according to physician-reported patient characteristics, and patient-reported data were matched according to patient-reported characteristics. The propensity score was estimated using a logistic regression model. Each patient in the non-switcher group was matched 1:1 to a patient in the switcher group, with replacement (i.e., the same patient being used as a match more than once) and allowing for ties (i.e., when switchers and non-switchers had identical propensity scores, so patients from one group are matched to more than one patient from the other group). For all variables, standard mean differences (SMDs) were within the – 10% and 10% limits, indicating good matching of the non-switcher and switcher groups on both patient and physician-reported characteristics. After matching, there were 160 non-switcher and 225 switcher patients with physician-reported data. The unmatched patient base was higher (n = 284) than that of the matched patient group (n = 225) because not all patients had data for each matching variable.

For all analyses conducted, a p value of < 0.05 was deemed statistically significant.

留言 (0)

沒有登入
gif