PERFUSE - A Real-World Study on Rheumatology Patients’ Persistence with Adalimumab Biosimilar SB5 in France: Impacts of Patient Experience

1. INTRODUCTION

Inflammatory Rheumatic Diseases (IRDs), such as Rheumatoid Arthritis (RA), Ankylosing Spondylitis (AS), and Psoriatic Arthritis (PsA) affect between 0.2 and 1.0% of the population, with RA being the most common of the three, accounting for a considerable clinical and socio- economic burden [1-5]. These pathologies progressively affect patients’ joints and bone structure, resulting in pain, reduced mobility, and difficulty in performing everyday tasks [6-8].

Treatments have evolved over the past decades, with the development of biologics designed to specifically target key factors in pathogenic mechanisms [9], such as Tumor Necrosis Factor (TNF), to limit the production of proinflammatory cytokines [10]. Since the first such drug (infliximab) [11, 12], other biologics have been developed and approved for commercial use, such as Adalimumab (ADL) [13]. As reference products came off patent, biosimilars were developed; these are now approved as equivalent to their reference biologics regarding their physicochemical properties, biological activity, and clinical efficacy, with similar safety and immunogenicity profiles [14]. ADL biosimilars were first made available for prescription in 2017 [15].

As biosimilars are more affordable than reference biologics, they enable the treatment of more patients within the same budget and thus have the potential to improve patient access to targeted therapies [16, 17]. Biosimilars may be perceived negatively compared to reference biologics, as their lower price is sometimes erroneously associated with lower quality [18]. Thus, physicians need to reassure their patients when changing their treatment to a biosimilar in order to reduce the risk of the nocebo effect (i.e., the appearance of a negative outcome due to an erroneous belief that the intervention will cause harm) [19, 20]. Indeed, providing a good patient experience is a key differentiating factor between clinically similar drugs [21]. This is why Patient-reported Outcome Measures (PROMs), which provide key information from the patients’ perspective [22, 23], are increasingly adopted and popular. PROMs can provide key insights into the patients’ perspective, e.g., RA care, daily life, treatment satisfaction, and well-being [24, 25], which are often unavailable in the case of conventional data collection methods (e.g. medical visits at investigation sites), generally resulting in the sparsity of available information [26].

Although clinical equivalence has been demonstrated in the development phase of biosimilars, new data on long-term safety and effectiveness, especially in indications other than the one studied in the registrational trial, serving the authorization process, are welcomed by authorities and clinicians.

The PERFUSE study was designed to address the need for real-world evidence involving patients routinely treated with SB2 or SB5 in France, being a longitudinal observational study on long-term persistence, effective- ness, and safety, designed with patient experience in mind, and including 914 patients who were receiving SB5, an EMA-approved ADL biosimilar, as routine therapy for their diagnosed rheumatic or Inflammatory Bowel Disease (IBD). Clinical data from routine specialist visits were reported by physicians for 12 months, and patient-reported data were collected using an online question- naire. To ensure their pertinence, questions were co-designed with patient associations.

In this analysis, the aim was to describe patient satisfaction of the rheumatology cohort treated with SB5 either as their first biologic therapy or switched from prior therapy, utilizing patient-reported data to assess their impact on 12-month persistence with SB5. We have presented the data reported by the patients concerning their illness, satisfaction, and confidence regarding their treatment, and their perception of their disease and medicine in general.

2. MATERIALS AND METHODS

3. RESULTS

3.2. Patient Training Practices

The frequency at which training was offered to patients was assessed for all ePRO respondents, including those who had not started SB5 at the time of completing the questionnaire. Attendance and satisfaction were only assessed among those who had started SB5 at the time of completing the questionnaire.

Respondent BSC-naive patients reported being offered training on the proper use of the SB5 injection device 74.8% of the time (n=80/107), with a 92.4% (n=73/79) attendance rate (Fig. 1b and 1c). Training mostly occurred at the hospital (94.5%, n=69/73). The length of each training was deemed satisfactory (neither too short nor too long) by 95.9% (n=70/73) of patients who attended the training (Fig. 1c). Training rates for naive patients were inconsistent, with sites offering training to 72.0% ± 35.3 (SD) of the patients on average, though individual site rates ranged between 0 and 100% (Fig. 1b).

Table 2.

Global Population	BSC Naive [N = 223 (43.9%)]	BSC Pretreated [N = 285 (56.1%)]	Total (N = 508)
Respondent population (completed ePRO and was confirmed to be treated by SB5 at the time), n (%)
	104 (46.6%)	117 (41.1%)	221 (43.5%)
Age in years, mean (SD)
Completed ePRO Did not complete ePRO	47.6 (13.5) 47.0 (15.1)	51.0 (12.8) 50.9 (15.0)	49.4 (13.2) 49.3 (15.1)
Disease duration, mean (SD)
Completed ePRO Did not complete ePRO	5.5 (7.7) 4.6 (6.8)	13.4 (11.1) 13.6 (11.8)	9.7 (10.4) 9.9 (10.9)
Number of women, n (% of each subgroup)
Completed ePRO Did not complete ePRO	64 (59.3%) 61 (52.6%)	63 (52.5%) 74 (44.8%)	127 (57.5%) 134 (47.7%)
Level of education of respondents, n (% of each subgroup)*
Less than high school diploma High school diploma or equivalent Up to bachelor’s More than bachelor’s	15 (14.4%) 37 (35.6%) 28 (26.9%) 24 (23.1%)	11 (9.4%) 53 (45.3%) 24 (20.5%) 29 (24.8%)	26 (11.8%) 90 (40.7%) 52 (23.5%) 53 (24.0%)

Note: BSC = biologic subcutaneous treatment; SD = standard deviation
* Data are available for the respondent population only; these data were collected from the ePRO.

Training on the proper use of the SB5 injection device was offered to 40% (n=48/120) of BSC pretreated patients, with an 82.6% (n=38/46) attendance rate (Fig. 1b and 1c). Training sessions mostly occurred at the hospital (92.1%, n=35). The length of each training was deemed satisfactory by 97.4% (n=37/38) of the patients (Fig. 1c). Training rates for BSC pretreated patients were also inconsistent with each site offering training to 31.8% ± 33.0% of their patients on average, though individual site rates were between 0 and 100% (Fig. 1b).

The rate at which training was offered to BSC naive patients was significantly higher than for BSC pretreated patients (p < 0.05) (Fig. 1b). Satisfaction with the training length was not significantly different between the groups. The rate at which BSC pretreated patients accepted to take part in a training session when offered was numerically lower than for naive patients, although not statistically significant.

3.6. Patient Perception of Illness

BSC naive patients had a slightly better view of their illness than BSC pretreated patients (B-IPQ scores of 47.6 ± 7.8 and 44.6 ± 10.2, respectively; p<0.05). However, scores remained moderate, indicating a generally ambivalent perception of their disease (Fig. 3c).

3.7. Association between PROMs and Persistence with SB5

Persistence with SB5 at 12 months was numerically lower for trained than for untrained patients [65.5%, 95% CI (54.9; 74.1) and 68.9%, 95% CI (58.9; 77.0), respectively; n.s.] (Fig. 4A). However, persistence with SB5 at 12 months was significantly higher in respondents than in non-respondents [66.4% (59.3; 72.5) and 48.7% (42.4; 54.8), respectively; p<0.05] (Fig. 4B).

A univariate Cox model analysis showed that patients who reported better satisfaction with the injection device were significantly less likely to stop treatment within the first year after SB5 initiation than those with lower satisfaction scores [HR=0.87 (0.79; 0.96), p<0.05]. This model also showed that patients with higher B-IPQ scores were more likely to stop treatment within this same period than those with lower scores [HR=1.03 (1.0;1.05), p<0.05]. Finally, no significant correlation with persistence was observed for PEPPI scores [HR=0.98 (0.95; 1.02), n.s.], BMQ general [HR=1.01 (0.97; 1.056), n.s.], or BMQ specific scores [HR=1.00 (0.94; 1.07), n.s.] (Table 3).

Table 3.

Score	N	HR (95% CI)	p-value
PEPPI score	221	0.982 (0.950; 1.015)	0.2774
Satisfaction with SB5 device	221	0.869 (0.791; 0.955)	0.0035*
BMQ specific	221	1.000 (0.938; 1.067)	0.9878
BMQ general	221	1.012 (0.969; 1.056)	0.6017
B-IPQ score	221	1.025 (1.000; 1.050)	0.0459*

Note: * Bold lines indicate a significant correlation. B-IPQ = Brief Illness Perception Questionnaire; BMQ = Beliefs about Medicine Questionnaire; PEPPI = Perceived Efficacy of Patient-Physician Interactions.

4. DISCUSSION

While most BSC naive patients were trained to use the injection device, most of the BSC pretreated patients were not. Numerically lower attendance rates for pretreated patients indicate that these patients felt training as not necessary, which has been reflected in higher confidence scores. While recommendations state that all patients receiving specific treatment for the first time should receive appropriate information and education on how to manage the new treatment [36, 37], we found specific training rates to be low overall. However, the reasons why training was not offered were not collected.

Results mirrored those observed in the gastro- enterology cohort, in which satisfaction levels and patient perceptions were similar. Indeed, these factors could be mainly related to the treatment itself, which was, however, the same for patients in both cohorts. However, results revealed contrasting training practices, with greater variability between rheumatology sites than between gastroenterology sites, and pretreated patients being trained more frequently by rheumatology sites than by gastroenterology sites. Additionally, the results of the correlation analysis were similar, with the key difference being the significance of the correlation between patient satisfaction and persistence [27].

Patient satisfaction with all assessed aspects of the prescription process was generally high, as has been observed in previous studies on biosimilars [38, 39]. Patients reported high levels of perceived efficacy in their dialogue with the prescribing physician and were generally highly satisfied with their experience at the time of prescription. No significant differences in terms of satisfaction were observed between BSC naive and BSC pretreated patients, nor between trained and untrained patients. This may indicate that neither prior experience nor specific patient training has a significant impact on any of the measured aspects of patient satisfaction when initiating or switching to SB5. It appears, therefore, that patient care at the time of the switch meets patients’ requirements without the need for separate specific training sessions. However, the reasons for switching to SB5, as well as the involvement of the patient in this decision, may constitute a potential bias in terms of treatment satisfaction and, consequently, persistence. Similarly, information on injection devices (i.e., whether a patient used syringes and auto-injector pens) was not collected. It is well-documented that auto-injector pens are generally better tolerated than syringes [40], which may have influenced patient satisfaction and experience.

Although patients’ beliefs about medicine in general were ambivalent, patients’ concerns about the potential adverse consequences of taking the study treatment outweighed the perceived necessity of using it to control their illness. This suggests that patient concern may principally stem from uncertainty surrounding the new treatment. Furthermore, patients’ beliefs about SB5 did not appear to be affected by their previous experience with a subcutaneous treatment. Indeed, pretreated patients had successfully been treated with a previous BSC and were thus less likely to have a negative preconception of SB5.

BSC naive patients had a significantly more positive view of their illness than BSC pretreated patients. Indeed, pretreated patients were living with their illness for much longer than BSC naive patients (Table 2) and were thus more likely to have experienced more flares. This is in line with previous studies, which have found disease duration to be negatively correlated with patients’ quality of life [41].

Patient training on the proper use of the injection device did not appear to be associated with persistence with SB5, either negatively or positively, and neither did BMQ nor PEPPI scores. These results seem to contradict previous findings that have reported enhanced communication strategies to lead to an increase in persistence [42, 43].

Respondents exhibited significantly higher persistence than non-respondents. This does not appear to be linked with patient status at inclusion, as the response rate was similar for both pretreated and naive patients, and, thus, may provide a useful tool to direct efforts to optimize persistence. This type of observation has been made in other therapeutic areas; cancer survival was observed to be better in patients who communicated via email with their physician [44], indicating that certain profiles of patients who are either more technologically inclined or more involved with their treatments may benefit more from them. Likewise, patients who are doing better may be more inclined to communicate.

However, although the association between B-IPQ score and persistence was statistically significant, the clinical significance was not clear. The analysis of factors affecting persistence was conducted using univariate regression, which may have introduced potential biases, such as limitations in capturing complex inter- relationships, confounding variables, and nonlinearity. Despite these limitations, the use of a multivariate analysis model was deemed unlikely to provide significant additional information compared to univariate analysis. This choice was justified by the observation that only one explanatory factor was clearly significant (satisfaction with the injection device; Table 3), while the remaining factors exhibited indications of multicollinearity.

Other limitations inherent to the design of the study and the type of data presented have been identified. First includes the relatively low response rate to the online questionnaire, which, at around 50%, was lower than we had hoped for, even though it appears to be in line with those observed in other studies using this delivery method (email along with email reminders) [45]. However, the reasons for non-response were unavailable; thus, the distribution of the participants was made by indication (RA, AS, or PsA), which may have led to potential biases. Second, large differences both in response and training rates were observed between the sites. To address this, reporting and training practices could be standardized to ensure training quality. However, this may be impossible for some investigation sites that do not have the necessary resources to provide proper patient training; this could possibly explain why some sites only trained a small percentage of patients or even none.

Additionally, pre-existing diagnoses, such as Juvenile Idiopathic Arthritis (JIA), which may have persisted to the adult age, may also affect patients’ experience with their ADL treatment according to some registries and surveys [46]. Patients with a primary diagnosis of JIA were excluded from this study in order to avoid this confusion factor [47]. However, the concomitant presence of other diseases that may be affected by ADL was not accounted for in the assessment of patient experience, as very few patients fell in this category (<1%).

The validated scores used in this study were chosen for their combination of simplicity (length and wording) and ease of interpretation. The bespoke satisfaction scores were constructed based on a commonly used scale for the evaluation of such measures. However, these questions have not been independently validated, and so their psychometric properties have not been assessed. Other possibilities for evaluating patient satisfaction exist, with the most common being 5 or 7-level Likert scales, as they are easily interpretable by the patient. We opted for a numerical scale, which could easily quantify the results. However, the interpretation of a numerical scale could be harder because patients’ understanding of the values within the scale may be different [48].

These ePRO results should be considered within the socio-economic context in which the data collection occurred, and which may have influenced patient-reported outcomes; this study was performed in France at a time when biologic and biosimilar treatments were at the forefront of debate in the health sector and patient associations were still working on providing accurate information about drug costs, funding, and reim- bursement. French patients have broad access to these treatments and are overall very well covered in terms of reimbursement for biologics. Thus, the economic benefits of biosimilar switching directly impact the healthcare system rather than the patients. The consensus among rheumatologists is that there is a need to promote the use of biosimilars in order to reduce expenses for the French public healthcare system [49], which makes the lack of training provided to pretreated patients all the more surprising.

CONCLUSION

This study has highlighted disparities in patient training practices, especially for pretreated patients, although this has not appeared to impact patient persistence with SB5 after 1 year. However, the study has indicated that patients who have expressed satisfaction with their treatment and those who have actively engaged in ePRO responses have tended to display greater levels of persistence. The satisfaction of patients could serve as a valuable marker to identify individuals more likely to exhibit non-persistence. Therefore, this finding suggests the need for standardized patient information practices across France, especially for pretreated patients.

REFERENCES