Treatment Satisfaction, Patient Preferences, and the Impact of Suboptimal Disease Control in Rheumatoid Arthritis Patients in Greece: Analysis of the Greek Cohort of SENSE study

Objectives: SENSE was an international, non-interventional cross-sectional study that assessed treatment satisfaction in patients with suboptimally controlled active rheumatoid arthritis (RA) who were under treatment with any approved agent exposed to ≤ 2 biological disease-modifying anti-rheumatic drugs (DMARDs) at the time of enrolment. The current publication concerns the subanalysis of the results from the Greek cohort. Methods: Treatment satisfaction was assessed with Treatment Satisfaction Questionnaire for Medication (TSQM), with good treatment satisfaction defined as TSQM global ≥80. Adherence to therapy was recorded on a visual analogue scale (VAS) and treatment expectations were assessed on a 7-point numerical rating scale. Results: Of 121 patients, 82.6% were women, of mean age 64.8 years and mean time from diagnosis 8.4 years. Patients had active disease (mean DAS28-ESR 4.5) and compromised functional status (mean [SD] HAQ-DI 1.1 [0.7]) while on treatment (43.8% on biologics and 5% on steroids). The mean TSQM global was 66.9. Treatment expectations were “general improvement of arthritis” and “less joint pain” (mean score [SD], 4.9 [1.8] each), “more joint flexibility” (4.8 [1.9]), and “lasting relief of RA symptoms” (4.8 [2.1]). Oral administration was preferred by 65.3% of patients. Good self-reported adherence (≥80%) was recorded in 93.4% of the patients. Treatment switch to another DMARD was planned by treating rheumatologist for only 49.6% of the participants, despite suboptimal RA control. Conclusion: Patients with suboptimally controlled RA in Greece have low treatment satisfaction and poor self-reported outcomes, albeit high self-reported treatment adherence. Similarly to the global SENSE study results, the need for patient-centric treatment approaches in order to improve disease outcomes is emphasised.


INTRODUCTION
Rheumatoid arthritis (RA) is a chronic, immune-mediated, inflammatory disease that, if not properly controlled, may result in progressive articular damage, loss of function, compromised quality of life, and increased mortality. 1 Two types of disease-modifying antirheumatic drugs (DMARDs), biological DMARDs (bDMARDs), and targeted synthetic DMARDs (tsDMARDs), are therapeutic options for patients with inadequate response to conventional synthetic DMARDs (csDMARDs) that are recommended by the European League Against Rheumatism (EULAR) for the management of RA. 2 Although the number of treatment options is steadily increasing and different drug classes have managed to slow down disease progres-sion, many RA patients remain suboptimally controlled 3 and sustained remission is rarely achieved. [3][4][5][6][7][8][9] It has been shown that patients who do not achieve treatment targets have worse short-and long-term outcomes and timely treatment adjustments according to treat-totarget (T2T) principles, considering patient preferences and perspectives are critical to prevent disability. [9][10][11][12] Although patients' perspectives are important determinants of treatment success in RA, they have not been adequately evaluated. Most of the studies on RA have focused on outcomes reported by the treating rheumatologists. Databases worldwide and local registries have contributed information on RA patients' and disease characteristics, including standard of care. Nevertheless, the evaluation of RA patients' preferences, expectations, and self-reported outcomes, such as adherence to treatment, particularly in suboptimally controlled patients with active disease, including patients with moderate to high disease activity despite treatment with DMARDs, can contribute valuable insights on potentially unmet needs and maximize treatment benefits. [13][14][15] Satisfaction correlates with patients' treatment expectations, which can differ from rheumatologists' treatment goals, 16 and is in turn linked to patient treatment adherence. 13,17,18 Increasing evidence suggests that adherence to RA treatment can be improved via patient support programs (PSPs) [19][20][21] and patient empowerment via access to digital health-related information for informed decision-making. [22][23][24] The effectiveness of the latter is related with the patients' digital health literacy (DHL), ie, the ability to access and use credible online health information. The international non-interventional cross-sectional SENSE study was conducted in 18 countries worldwide between September 2018 and May 2019 to determine the impact of inadequate response to DMARDs on treatment satisfaction and various disease outcomes and to analyse patients' attitudes and perspectives toward treatment and their disease. 15 SENSE also provided an opportunity to assess DHL in a large multinational cohort of patients with RA. In Greece, local RA databases, including the Hellenic Registry of Biologic Therapies, the nation-wide e-prescription platform, and the more recent country-wide database created by the RA Working Group of the Hellenic Rheumatology Society, have contributed information on RA and afflicted patient characteristics. [25][26][27][28][29][30][31][32][33] Here, we report a sub-analysis of the global SENSE results from the patients that have been enrolled in seven rheumatology centres (public and private hospitals) in Greece (Supplementary Table 1

Study design
The SENSE study was performed according to the Declaration of Helsinki with prior approval from each site's Scientific Committee. Patient selection criteria included the following: Diagnosis of RA using either the 1987 revised American College of Rheumatology (ACR) or the 2010 ACR/EULAR classification criteria for RA; ongoing treatment with any approved csDMARD, tsDMARD, or bDMARD; and exposure to ≤2 bDMARDs at the time of the enrolment. All patients had to have residual disease activity as measured by Disease Activity Score, 28 joints (DAS28 >3.2) for 1 to <4 months before enrolment despite having received the full tolerable dose of current DMARD therapy for ≥3 months. Consecutive patients attending a routine rheumatologist office visit and fulfilling enrolment criteria were included in the study. Physicians collected data during a single scheduled visit.

Assessments
Clinical parameters and socio-demographic characteristics were collected for all patients. Medical history including comorbidities (coded via the Medical Dictionary for Regulatory Activities system organ class level) and concurrent treatment, both for RA and overall were collected. Past medications for RA were also collected.
Physicians were asked to report if switch to a different DMARD was planned for their patient, and the mode of action of planned treatment switches was captured. The primary objective of the study was to assess patients' treatment satisfaction related to current RA treatment using the Treatment Satisfaction Questionnaire for Medication, version 1.4 (TSQM v 1.4). 34 This tool incorporates Effectiveness, Side Effects, Convenience, and Global Satisfaction domains, with scores ranging from 0 (poor satisfaction) to 100 (perfect satisfaction). Good treatment satisfaction is defined as TSQM global ≥80. 35 VAS using numeric rating scales (NRS) were used to assess morning stiffness severity and duration (in minutes) as well as pain in the past 7 days (range 0 = "no stiffness/ pain" to 10 = "worst possible stiffness/pain"). 15 The following validated patient-reported outcomes (PROs) were used: Health Assessment Questionnaire -Disability Index (HAQ-DI) for physical function, Functional Assessment of Chronic Illness Therapy -Fatigue (FACIT-F) for fatigue, Work Productivity and Activity Impairment -Rheumatoid Arthritis (WPAI-RA) for workability, Short Form 36 Health Survey Questionnaire (SF-36) physical and mental component score for health-related quality of life (HRQoL). [36][37][38][39] Self-reported adherence to medication was assessed using VAS, with good adherence defined as ≥80% on VAS. 40 Patient medication preference information (PMPI), including preference for route of administration, combination therapy, time to effect, and acceptable side effects, was assessed by a 6-item questionnaire developed by AbbVie (Supplementary Table 2). 15,41,42 Patient expectations for pharmacologic treatment were assessed using an 11-item questionnaire with a 7-point NRS (1 = "no improvement needed" to 7 = "the most improvement needed"). The need for PSP was assessed using a 17item questionnaire with a 7-point NRS (1 = "not needed at all" to 7 = "very much needed"). Healthcare resource utilization (HRU) during the three months before enrolment was also recorded and used to determine HRU over the past 12 months (by multiplying 3-previous month data with 4). DHL was assessed by eHEALS, a self-report tool of 10 questions based on individuals' perceptions of their skills and knowledge within each measured domain, providing scores ranging from 8 to 40; a higher total eHEALS indicates greater perceived skills at using online health information to help solve health problems; a score of <26 was considered to represent poor digital health literacy (DHL). 24,43 Statistical analysis All statistical analyses were carried out using SAS ® software (version 9.4; SAS Institute, Cary, NC, USA). Quantitative data were described by the statistical parameters valid N, missing N, mean, standard deviation, median, minimum, maximum, lower quartile (25%), and upper quartile (75%). Qualitative data were described with (absolute and relative) frequency distributions. Twosided 95% confidence intervals (CIs) were calculated when appropriate. Descriptive statistics using the full analysis set (FAS), which included all patients who fulfilled all inclusion criteria, was employed, without data imputation. All results reported are based on the number of FAS patients, unless otherwise specified. The sample size calculation of the global study was based on standard deviation information on Global Satisfaction measured by TSQM v1.4. A sample size of n=1500 was expected to be able to provide a 95% CI with a half width of 1.01 in the overall study population. 15,34 For country-specific analysis, it has been estimated that a sample of n=30 -200 will be able to provide a 95% CI with a half width of 7.47 to 2.79. Subgroup comparisons of patients with or without any comorbidities were conducted to identify any differences in PMPI, expectations and PROs. For continuous variables, Wilcoxon-Mann-Whitney tests used; for categorical variables, chi-squared tests or exact Fisher tests were used.

Clinical parameters and sociodemographic characteristics
A total of 121 patients were enrolled in SENSE study in Greece and were included in the full analysis set (FAS).

TITLE TREATMENT SATISFACTION, PATIENT PREFERENCES, AND THE IMPACT OF SUBOPTIMAL DISEASE CONTROL IN RHEUMATOID ARTHRITIS PATIENTS IN GREECE: ANALYSIS OF THE GREEK COHORT OF SENSE STUDY
Demographic characteristics, employment status, and level of education are described in  Table  3). The mean (SD) number of comorbidities was 2.7 (2.1), with the most frequent being cardiovascular comorbidities (55.4%) followed by metabolic and nutrition disorders (43%), endocrine disorders (34.7%), musculoskeletal and connective tissue diseases (24.8%), and psychiatric disorders (24.8%). HRU was high and a previous medical visit for RA was reported by 82.6% of the patients, all of which were on an outpatient basis. The mean (SD) number of visits was from 2.1 (1.1) to 8.5 (4.5) during the previous 3-and 12-month periods prior to enrolment, respectively.

Medication and Treatment Strategy
The most frequently used RA medications included csDMARDs, namely methotrexate (62.8%), hydroxychloroquine (17.4%), and leflunomide (11.6 %); only 5% of patients were treated with systemic corticosteroids at the time of evaluation ( Table 4). Among all patients, 43.8% had been treated with bDMARDs; 45.3% of these patients were on monotherapy. Interestingly, despite long-standing disease and suboptimal symptom control   (18.9). The low level of satisfaction was driven by the low effectiveness subs core, in alignment with the suboptimally controlled, active RA (Figure 1).

Secondary Outcomes
RA affected productivity, functional status and overall QoL ( Table 5). Good self-perceived adherence, defined

Patient Expectations for Pharmacological Treatment
The highest-rated treatment expectations were general improvement of arthritis, less joint pain, lasting relief of RA symptoms, more joint flexibility, and less joint swelling (Figure 2).

PSP participation
In terms of need for patient support, patients assigned the greatest importance to having access to educational material that focused on RA disease and therapy as well as to a call centre and a starter pack with all information about the patient-support programs.   Digital health literacy Based on eHEALS score, the majority of patients were found to have poor familiarity with digital tools for the management of their disease. The mean (SD) patient score was 15.4 (8.8), and the highest patient score was 32. In general, more than half of the participants disagreed or strongly disagreed with the following statements: "I have knowledge of the health resources that are available on the internet", "I know where and how to find helpful health resources", "I know how to use the internet to answer my health questions", and "I know how to use the health information found on the internet". Only 12.8 % of patients of the 109 available responses agreed or strongly agreed that they can differentiate high-quality from low-quality healthcare resources on the internet, with 6.4% of patients agreed or strongly agreed that they felt confident in using information from the internet to make healthcare decisions.

DISCUSSION
This study aimed to assess the real-world perspective and treatment expectations of patients with suboptimally controlled RA, information that is considerably underrepresented in the literature. It has been previously shown that patients' and physicians' perceptions of RA-related treatment priorities and disease activity may differ. 16,44,45 The main findings of this subanalysis from the Greek RA patients and the overall SENSE results demonstrated that despite the low level of satisfaction, as determined via TSQM global score, the vast majority of patients had high self-reported adherence to therapy (adherence ≥80%). This finding should be further explored and confirmed in other cohorts. In the Greek subanalysis, both the TSQM global score and good self-reported adherence to treatment slightly exceeded the overall study results. Conversely, to the global SENSE analysis, both the mean TSQM Global Satisfaction and Effectiveness domain subscores were among the lowest, whereas the safety (Side Effects) domain subscore was the highest. The physical function, and overall performance and HRQoL of the patients were negatively affected by the ongoing disease activity. These data further support the concept of the T2T strategy aiming at remission or low disease activity, since better disease's control might improve patient satisfaction. Patients' treatment expectations are associated mainly with control of the disease overall and on specific RArelated symptoms, such as joint pain, swelling, fatigue, and stiffness. Our results suggest a preference for oral versus parenteral therapy among the majority of patients, with about a third not favouring combination therapies, and also a preference for drugs with a rapid onset of action. These results comply with the overall study results. Interestingly, despite suboptimal disease control and long-standing disease, treatment switch to another DMARD was planned by the physicians for only approximately half of the participants in both the global and the present country-specific analysis of SENSE study. The analysis of the global dataset showed that lower patient global satisfaction scores were associated with planned treatment switches. 15,46 Our country-specific subanalysis did not assess the willingness of patients to receive treatment intensification and, therefore, could not evaluate whether suboptimal patient satisfaction is associated with patient acceptance of treatment changes or vice versa, whether good patient satisfaction despite poor disease control prevents rheumatologists from treatment adjustments. A correlation between therapeutic satisfaction and the patient's attitude towards treatment has been described, 47,48 and a recent study has shown that patients who report treatment satisfaction exhibit a weaker inclination to accept treatment intensification, regardless of their DAS28 score and duration of disease. 49 The data from the SENSE study further corroborate results from other Greek and international studies showing an inconsistency between the treatment recommendations for T2T and clinical practice. The results of a Greek study of patients in the early stages of arthritis similarly showed that only 62.4% of participants who experienced medium or high disease activity after 6 months of treatment were subject to treatment adjustments. The implementation of treatment modifications was reportedly followed by a significant decrease in disease activity after 2 years. 50 Likewise, in a recent multinational observational study, the T2T guidelines were appropriately applied in only 59% of patient visits. 51 We believe that the rheumatological community needs to consider carefully these findings to identify the specific barriers of the clinical implementation of T2T concept. Comparable therapeutic inertia, defined as "the failure to initiate or intensify therapy in a timely manner, according to evidence-based clinical guidelines", is certainly present in the treatment of other chronic diseases. 52 As literature  TITLE   TREATMENT SATISFACTION, PATIENT PREFERENCES, AND THE IMPACT OF SUBOPTIMAL DISEASE CONTROL  IN RHEUMATOID ARTHRITIS PATIENTS IN GREECE: ANALYSIS OF THE GREEK COHORT OF SENSE STUDY shows, the potential discordance between physicians and their patients regarding treatment target definition, disease perception and need for treatment adjustment can significantly affect therapeutic decisions in patients with suboptimal disease control, though evaluating the discordance was not the purpose of the study. 53 Comorbidities in RA are common and have a negative effect on patient functioning, morbidity, and mortality. 4 Similarly to the overall study results, comorbidities were encountered in the vast majority of the patients from Greece, and 82.6% reported receiving medications for other diseases, with the mean number of drugs administered being 2.1. There was an overlap in the most frequently reported comorbidity categories between the Greek cohort and the overall study population. Nevertheless, except for musculoskeletal/connective tissue disorders, for which the incidence was comparable in the present cohort and overall study population, the incidence of cardiac, metabolic/nutrition disorders, endocrine and psychiatric disorders was higher in the Greek patients. Interestingly, the incidence of psychiatric disorders was 3-fold higher in this subanalysis. It is worth noting that the presence of a comorbid disease correlated with worse disability (HAQ-DI) and lower TSQM global satisfaction scores. These findings further support the importance of the effect of comorbidities on the outcome of RA and the necessity for their effective management. Additionally, patients had poor digital health literacy, and, therefore, poor familiarisation with tools for the management of their disease. Concerning the benefits of digital health resources, the patients reported that their highest prioritization was for receiving information on general RA disease-and medication-related topics through a PSP program and their lowest for digital lifestyle interventions, such as social media, smartphone, and website contents. The eHEALS study results revealed low DHL, highlighting the need to develop health promotional programs addressing DHL and digital tools tailored to the needs and pragmatic capabilities of the RA population. New information-and communication-technologies may substantially contribute in a more accurate monitoring of disease-related parameters while offering much-needed patient education. 54 As the RA population gradually shifts towards patients with a higher degree of familiarity with digital content and applications, these educational activities could be further developed and applied to a larger group of patients. Except for the prevalence of females over males, there were differences in the sociodemographic characteristics of this subanalysis and the global SENSE results. Some of these differences, particularly in occupational status, are attributed to the age range of the participants. Thus, based on mean age, the Greek cohort patients were slightly older (mean age of overall study patients 58.4 years old), which in turn accounts for the higher percentage of retired patients in this subanalysis. The observed differences in the incidence of comorbidities between this subanalysis and the overall SENSE results are likely to be attributed to the older age of the current patients. A comparable percentage of patients in this subanalysis and the overall SENSE results had university education. Psychosocial factors, such as education and occupational status as well as demographics, amongst other factors, are likely to influence and account for the potential differences, albeit small, in patient expectations and preferences, DHL and PROs in this subanalysis and the overall SENSE results. Similarly to this subanalysis, csDMARDs were the most frequently prescribed medications. Differences between individual bDMARD prescriptions in this subanalysis and the global SENSE results can be attributed to local therapeutic protocols and potentially reimbursement policies in the participating countries.
Concerning the limitations of the study, by design, noninterventional studies hold certain limitations, such as selection and recall bias and lack of a control group. The focus on a specific patient group with suboptimal disease control may limit the generalizability of our results to all RA patients. Although PROs reflect subjective patient assessments, however, this effect was counterbalanced by the use of validated PRO tools. Similarly, VAS for the determination of self-reported treatment adherence is validated and highly correlates with electronic monitoring results in patients with chronic conditions, including RA. 40,41,55 No validated questionnaires were available for assessing the need for PSP, treatment preferences, and expectations. The imbalance in the sizes of groups with and without comorbidities as well as the presence of potential confounding factors warrant caution when interpreting the results of subgroup analysis. These results, therefore, need to be confirmed by using validated measures in future studies. Because of the size of the Greek sample, further subanalyses and correlations to specific outcomes were not possible.
This study provides an in-depth understanding of patient needs and perspectives, also identifying unmet requirements for treatment adjustments that will align with recent therapeutic standards and the T2T principles. self-reported outcomes, albeit a high self-reported treatment adherence. These data further support both the value of treatment approaches targeting to abrogation of inflammation and emphasise the need of documenting patients' perspectives to improve disease outcomes.

AUTHOR CONTRIBUTIONS
All participating authors contributed equally to the gathering of information and writing and reviewing of the article. 33