Rheumatoid arthritis (RA) is a systemic autoimmune disease that requires early diagnosis and treatment to improve symptoms, reduce structural damage and control functional and psychological involvement. The different, both diagnostic and therapeutic, strategies that are currently used, such as tight control or treat to target, aim to achieve these objectives. Moreover, it has been demonstrated that an early approach also reduces costs, thus improving therapeutic efficiency.1 However, although most studies take good patient compliance for granted, other studies show that adherence to RA treatment is inadequate and its effectiveness can be diminished.2 Therefore it is essential to detect the causes impeding therapeutic adherence, and to implement methods to improve it.

Correct therapeutic compliance can be made difficult for many reasons.3,4 The elements of therapy are highly complex and can be affected by different factors, associated with the patients and their disease, the health professionals and the area where they work, and with the drug used and its possible side effects.5,6 A Spanish study suggests that young men have the poorest adherence.7 It is calculated that between 30% and 80% of patients with RA and other chronic illnesses fail to follow their treatment appropriately.8,9

It has been demonstrated that follow-up of patients with RA must be rigorous with frequent visits structured according to established therapeutic protocols and programmes.10 Current routine assessments are increasingly more thorough and cover: disease activity and its relationship with quality of life, functional capacity, comorbidities, disease progression, efficacy of treatment and its possible adverse events (AE).1 However, in the care of the chronic patient, policies to assess levels of treatment compliance and reasons for it are not often implemented, although there are several strategies for doing so.11,12

The synthetic or conventional disease modifying antirheumatic drugs (DMARDs) are the first step in treating RA, and are often accompanied by AE of varying intensity.5,13 In fact, more than half of RA patients refer to the AE as the main disadvantage of the synthetic DMARDs, hence the consequent lack of adherence.14 However, there are often other reasons and concerns that must also be considered.13

For these reasons, the Treatment Adherence Observatory, in collaboration with the Spanish Rheumatology Society, launched the OBSERVAR study, with the main aim of discovering the reasons why RA patients in Spain are failing to adhere to treatment with the synthetic DMARDs.

The study started in January 2014 with a systematic literature review referring to treatment adherence using PubMed and the Google Scholar search engine based on scientific publications and other trustworthy sites, and we found 66 claims with a low level of evidence (III or IV according to NHMRC, National Health and Medical Research Council, criteria),15 divided into 3 subject blocks or categories: 24 patient-related reasons, 23 health professional-related reasons, and 19 treatment-related reasons for failing to adhere to treatment.7,16–18

Nineteen rheumatologists were invited to take part, selected by a scientific committee, seeking to represent the greatest number possible of autonomous regions, and cover both large and small hospitals.

After contacting all the participants by telephone, the list of selected claims were circulated twice to the participants’ personal email addresses. The first questionnaire was sent in July 2014 to the 19 rheumatologists selected, and again one month later to the 18 participants who responded to the first mailing (see attached appendix).

The Delphi process that we applied used validated scientific methodology to enable agreements to be reached, and each participant indicated their level of agreement with the claims using a Likert-type psychometric scale, scoring between 1 (minimum) and 9 (maximum acceptance) following the RAND/UCLA system (RAND corporation and University of California in Los Angeles).19

We then performed a descriptive analysis of the 66 reasons selected using the statistical tools of Microsoft Excel 2012 to determine: mean and standard deviation (SD); median and interquartile range (IQR); mode, minimum value, maximum value, and coefficient of variation (COV).

Two different criteria were established to assess the final level of agreement:

1. Criterion based on the group median: although the mean is routinely the most frequently and arithmetically convenient parameter used, we considered the median a better measure of central tendency, since it is less sensitive to the extreme values in asymmetric distributions.

Based on the median, 1–2 and 8–9 were considered disagreement and very consistent agreement respectively; 3 and 7 disagreement and consistent agreement; and the intermediate values (4.5 and 6), completely inconsistent.

2. Criterion based on simultaneous compliance: given that SD, IQR and COV represent the dispersion with respect to the central value, we considered unequivocal consistency of agreement when the mean and the median were ≥7, the SD and the IQR were ≤1.00, and the COV was ≤.25.

The results obtained were crossed with the aspects of segmentation of the participants, the autonomous region where they worked, and the number of RA clinical histories, applying the Students t-test and the χ2 test.

The study was approved by the clinical research committees of the referral hospital “A Coruña University Hospital Complex”, and then by the rest of the participating hospitals.

The project was carried out from March to September 2014. Eighteen of the 19 participating rheumatologists (94.74%) completed the 2 circulations of the Delphi process. This included 13 autonomous regions (Andalusia, Catalonia, Madrid, Valencian Community, Galicia, Castilla-León, Basque Country, Castilla-La Mancha, The Canary Islands, The Balearic Islands, Extremadura, Asturias and Cantabria), and Melilla as an autonomous city.

Table 1 shows the patient-related reasons for failing to adhere to the synthetic DMARDs. We must highlight that among the 24 related variables, the level of agreement based on the median was very consistent in 7 (29.10%), consistent in 11 (45.83%), and inconsistent in 6 (25%). There was no disagreement in this section. Based on the simultaneous criterion, there was only consistent agreement in one statement (4.17%).

Table 2 specifies that, among the 23 health professional-related reasons, the level of agreement according to the median was very consistent in 3 (13.04%), consistent in 18 (78.26%) and inconsistent in 2 (8.70%). There were no disagreements in this section either. When the simultaneous criterion was analysed, consistency was found in only 2 agreements (8.70%).

After assessing the 19 reasons for a lack of adherence to the synthetic DMARDs relating to the “treatment itself”, the level of agreement was consistent in 11 (57.89%), there was only one disagreement (5.26%) and 7 inconsistencies (37.84%) considering the criterion of the group median. There was no agreement based on the simultaneous criterion (Table 3).

In this study, designed to understand the main causes for the lack of adherence to synthetic or conventional DMARDs in the treatment of RA, only 3 reasons stood out as very consistent: (1) not knowing what to do in the event of an AE; (2) not having undergone adherence screening by health personnel; and (3) not having applied the strategies that improve therapeutic adherence.

The therapeutic adherence concept was used as a synonym for therapeutic compliance, defined as the extent that a patient's behaviour, in relation to taking their medication, following a diet or changing lifestyle habits, matches the recommendations agreed with healthcare staff, or the extent that a patient acts in accordance with the dose, dosing regimen and prescribed times, which must be distinguished from persistence of treatment, which is the number of days of continuous usage of the medication over a specific time.8,9 Another similar term, although not identical, is therapeutic concordance, a concept that is based on a negotiation between clinicians and patients, as equals, and opens the way for the latter to differ from the former in evaluating the benefit-risk balance of the drugs.19 All the variables analysed in this study only refer to therapeutic adherence or compliance.

A strict approach to RA, both from a diagnostic and a therapeutic perspective, clearly improves the progression of the disease and its prognosis.9,13 However, very often, the therapeutic indications are not clearly followed, and the results and objectives set are not as desired, resulting in diminished effectiveness.20 Control of pharmaceutical expenditure is optimised if medical prescriptions are strictly followed, and consequently, inadequate adherence reduces therapeutic efficiency.14,21 The main reasons for poor adherence and reduced therapeutic persistence have been widely studied.8,21,22

This study analyses the agreements reached by a panel of rheumatologists chosen to represent the Spanish rheumatological community, randomly distributed throughout the autonomous regions. The hospitals selected were tertiary and regional; we considered they would provide a significant sample of the opinions of Spanish rheumatologists, and that the consensus reached faithfully illustrates the reality in our setting regarding therapeutic adherence in RA. The Delphi method is widely generalised and validated and its results should faithfully reflect the reality surrounding us.18,19 However it is possible, since the results come from subjective opinions, that there could be some bias, which is obviously reduced by the methodology itself and the consensus reached over several stages. We only obtained national results, which therefore cannot be extrapolated to other countries or scenarios.

Poor therapeutic adherence is not specific to RA, it also occurs with other rheumatic diseases such as systemic lupus erythematosus, psoriatic arthritis and ankylosing spondylitis, as well as diseases of other specialities such as infectious, neoplastic, cardiovascular, metabolic, and possibly most chronic diseases.10,21,23,24

Overall, we found agreement for 75.8% of the claims surveyed. There was only one disagreement regarding the treatment-related reasons for lack of adherence: “imprecise or not appropriately reassessed diagnosis during follow-up”. On the other hand, categorical agreement was reached for three of the claims: “not knowing what to do in the event of an adverse event with the DMARD”, “potential non-compliant patients not having undergone adherence screening by health personnel”, and “not having applied strategies to improve therapeutic adherence”. The only disagreement with regard to the literature concerned an imprecise diagnosis as a cause, since something that could occur in other chronic illnesses, seems less likely in RA, where there are specific programmes for early diagnosis and a differential diagnosis with other diseases that simulate the clinical characteristics of RA, and follow-up protocols that repeatedly confirm diagnosis and improve the comprehensive approach to the disease.23,25 The 3 reasons for which there was categorical agreement were: improving the doctor-patient relationship, indicating different aspects of a failure to convey information (not knowing what to do in the event of an AE), a failure to detect the problem (lacking adherence screening methods) and lack of action (failing to use interventions to improve therapeutic adherence).

The patient-related reasons for which there was the greatest consistency were grouped under two sections: “reasons relating to an insufficient level of knowledge of the disease, treatment and the importance of adhering to it” and “a lack of knowledge of the repercussions of their disease and therapeutic non-compliance”. Believing that improvement with the use of a biological DMARD rendered a conventional synthetic DMARD unnecessary and stopping treatment after unexpected AE because of not knowing what to do, would also be included in this section. On the other hand, the other aspect was internal in nature: “lack of commitment and responsibility in complying with the prescribed treatment”; “not knowing that they are really ill”; “overestimating one's own adherence profile”; “not understanding and not asking”, “not sharing the same aims as the doctor”, “not taking part in decision-making”.

Each treatment is really personalised, re-evaluated and adjusted if necessary at each follow-up; but just as important as doing so correctly is that the patient should feel that they are receiving really “personalised treatment”, controlled, with specific objectives, that they are appropriately instructed about their illness, the importance of adherence and consequences, thus motivating and reinforcing their efforts.1 The patient-related variables indicate that medical and nursing information must improve, since often the professionals agree that patients have insufficient knowledge of the need to adhere to their treatment, or lack the necessary involvement to understand the danger of discontinuing it, or the benefits of continuous treatment with synthetic DMARDs for patients receiving a biological DMARD to improve their effectiveness and reduce immunogenicity.

In this regard, there might be differences between the different synthetic DMARDs and also between the biological DMARDs, concomitant medication with MTX having been studied more than with other conventional DMARDs.3,26 There are also differences with other entities such as psoriatic arthritis or ankylosing spondylitis where monotherapy with biological DMARDs is more frequently accepted and useful.3 Most of the variables analysed can be improved by providing patients with RA health education plans, and implementing rheumatology nursing clinics.12

This study only analysed the synthetic DMARDS, MTX being the most widely used. Although there is literature on both the synthetic DMARDS and the biological DMARDs, there are no major differences in terms of therapeutic adherence or reasons for discontinuation. Low social class was highlighted among the data associated with poorer adherence; this is associated with poorer understanding of the disease and its treatments, and also higher disease activity.

In general, acknowledged among the health personnel-related reasons were “a low awareness on this point among the professionals themselves”, resulting in insufficient follow-up of this aspect, lack of recording it, leaving it out in guidelines and recommendations, etc., and “not using methods to detect non-compliant patients”, “not applying interventions to improve therapeutic adherence”, “not undertaking awareness-raising programmes among healthcare personnel” and even “not relating poor adherence with potential inappropriate response” were highlighted. Again, all the programmes aimed at raising the awareness of health staff could improve the perception of this issue and, once again, providing patients with nursing clinics and information clearly improves awareness of the importance of maintaining the drug to achieve relevant clinical results. However, it is possible that some of the variables analysed relate more to the health structure itself than to the activity of healthcare staff.

With regard to “the treatment itself” as the source of non-compliance, rather general aspects were admitted, that might refer to any chronic disease: “polymedication”, “problems synchronising medication intake”, “lack of reminder techniques on containers”, etc. The data is contradictory on whether adherence is better if the drug is administered orally or parenterally, these results occur again with the biological DMARDS where, contrary to what one would imagine, therapeutic adherence with subcutaneous administration is also not as desired.27,28 In this regard, adherence could be much improved by administering the treatments intravenously, because control is stricter as this must be done in hospital. Although one study indicates that the therapeutic adherence in this subgroup of patients was not as expected either, depending on the combination or otherwise with a synthetic DMARD.29

We appreciate that the AE of the conventional DMARDs are considered a lesser cause of lack of adherence, some aspects are even inconsistent, with values that would bring them close to disagreement. This leads us to believe that, since differences were certainly shown between the doctor and their patients in terms of the appreciation of AE and their importance, patients might not be disclosing all their AE. However, different studies show that the AE of both the synthetic and the biological DMARDs are a frequent cause of reduced therapeutic adherence.4,21,29 In this regard, this subject should be approached from the perspective of an adverse effect as such, or a fear that one will occur, and on the other hand actions to encourage awareness4 among doctors as well as patients should be promoted, which would result in early detection of AE and agreed action after their onset, since they should not always involve discontinuing the drug.

Our study has some limitations. Firstly, although the number of rheumatologists consulted might appear small, it was considered sufficient after evaluation by an epidemiological/statistical advisory team. Similarly, the level of health care was heterogeneous with specialists that visit a very different number of patients in hospitals of different care levels. However, this enabled us to cover a large and varied group of rheumatologists with different clinical practice, which seems more a pro than a con. We did not include graduates of nursing or other healthcare professions in the study either, which might also reduce the importance of some of the results obtained. Furthermore, given the great individual variability of the reasons for therapeutic noncompliance, it would have been interesting to define more restricted, specific populations such as: patients with synthetic DMARDs combined with subcutaneous or intravenous biological DMARDS, the treatment time with the different drugs, etc. However, and despite all the aforementioned limitations, we believe that this study is valid and that the results obtained are consistent.

Based on this experience, we consider it essential to inform patients sufficiently about each new treatment introduced and its possible AE, to include an assessment of each patient's therapeutic compliance profile in the clinical routine, and to reinforce patients’ motivation for therapeutic compliance throughout their follow-up.

Roche Pharma España.

The authors have no conflict of interest to declare.