The prognosis of chronic inflammatory diseases such as rheumatoid arthritis (RA) has improved substantially in the last decade with the introduction of biological therapies, but despite the demonstrated efficacy and effectiveness of these treatments, toxicity is also an important factor to take into account.1 Up to 20%–40% of patients treated with methotrexate (MTX) have been observed to experience tolerability problems requiring drug withdrawal. Adverse events (AEs) are among the factors that most limit the use of some treatments and determine the management of patients in daily clinical practice.2

Many studies analyze the tolerability of pharmacological therapies for RA, but traditionally it is examined only from a clinical point of view and in overall terms (using treatment drop-out/discontinuation rates, persistence rates, laboratory tests or brief interviews with patients during a visit), without taking into account the impact of symptom or events treatment-related on patient's life. From the patient's perspective, tolerability includes detection of treatment-related events, but also assessment of their impact on daily life. For example, fatigue may not be a very important or critical AE from a clinical perspective, but it may cause a high degree of disability from the patient's point of view in terms of professional activities. These patient-reported outcomes (PROs) are becoming increasingly essential in clinical trials conducted in patients with chronic diseases, observational studies and, ultimately, daily clinical practice. PROs include a part of the patient's life that is not assessed at medical visits and may provide information that affects therapeutic decisions, such as suspension of a drug.3–5

Some studies have noted that, in general, clinicians report a substantially lower number of AEs than patients do,6,7 and that patients tend to provide a more detailed description, including information about its impact on, for example, health-related quality of life (HRQoL).8 In several diseases such as schizophrenia and bipolar disorder, specific questionnaires to assess the impact of secondary effects on HRQoL have begun to be validated and used.9,10 The few studies on RA that include the patient's assessment of tolerability do so using non-standardized tools with ad-hoc questions for patients or with HRQoL questionnaires that are not designed to assess tolerability.11,12 The patient's perspective is a concept that is gaining such importance that some countries such as the United Kingdom, the United States, Australia and Sweden have added reports supplied by patients to their pharmacovigilance systems.13,14

This fact, as well as the lack of validated tools, arouses the interest of authors to conduct a study with the objective of designing and validating a questionnaire to assess tolerability in patients under treatment for RA (TOL-AR-18) and to describing the potential utility of the questionnaire to assess tolerability of treatment for rheumatoid arthritis in the daily clinical practice of rheumatologists and other professionals involved in the management of these patients. The results of the validation are presented below.

The study sample included 66 patients with socio-demographics and clinical characteristics as described in Table 2. AEs reported by patients in the TOL-AR-18 were assessed by clinicians like mild/moderate and not serious. Patients showed less tolerability, meaning higher scores facing all AEs compared with the clinician's assessment in all dimensions (P<.05) as seen in Table 3. The most common event present were hair loss and tiredness/fatigue (Table 4). Pain, tiredness and general discomfort caused by treatment, affected more than 15% of patients very much or quite a lot (Table 5). Clinicians have to ask for other possible AEs, but no AEs apart from those included in the questionnaire occurred at the time of the study. Regarding tolerability, patients had an overall mean (SD) score on the TOL-AR-18 of 29.8 (7.0). The dimensions a little more affected were ‘general health’ and ‘psychological–cognitive’, although all dimensions had very low mean scores (Table 6).

TOL-AR-18 Validation (Psychometric Properties)

The psychometric properties assessed in the validation of the tolerability questionnaire are described below.

The rate of non-response to the 18 items was 1.5%. Reliability of the TOL-AR-18 questionnaire was assessed in terms of internal consistency with, Cronbach's α values greater than 0.95 for each item, which indicates that all items were very inter-related. Correlation between each item and overall score was greater than 0.3 that is a satisfactory correlation.

To asses the construct validity, TOL-AR-18 scores were compared with DAS-28 and with number of AEs. Patients with moderate/high activity had higher scores (lower tolerability) (Fig. 1), with statistically significant differences when compare with all other patients (P<.05). This significance was due to patients in remission who have very low scores. Statistically significant differences were also found for patients without AEs compared to all other patients. Patients with ≥2 AEs showed higher scores. To asses the construct validity, also Spearman's correlation coefficient (Rho) between TOL-AR-18 dimensions and ARTS questionnaire scores (items 9, 10 and tolerability dimension) was calculated. The dimensions of the questionnaire TOL-AR-18 had an inverse and low-moderate correlation with ARTS questionnaire (item 9 and item 10). Item 10 of ARTS had higher correlation coefficients with the dimensions of the TOL-AR-18 than item 9: between 0.21 and 0.38 and between 0.09 and 0.34 respectively. ‘General health status’ and ‘psychological/cognitive’ dimensions of TOL-AR-18 showed the highest correlation coefficients with tolerability dimension of the ARTS while ‘energy-vitality’ showed the lowest (Table 7). In addition, in the analysis of answers to the question about general health status, differences in the ‘daily activity’ and ‘psychological–cognitive’ dimensions and overall score for the TOL-AR-18 were found to be statistically significant (P=.008; 0.024; 0.041). Patients who assessed their general health status as ‘somewhat poor’ obtained a higher score (lower tolerability) on this questionnaire.

Fig. 1.

Result of the TOL-AR-18 scores according to disease activity (DAS-28).

(0.12MB).

Table 7.

Correlation Between TOL-AR-18 Dimensions and ARTS Questionnaire (Items 9, 10 and Tolerability Dimension).

TOL-AR-18 dimensions	ARTS
	Item 9	Item 10	Tolerability dimension
Daily life	Rho=−0.23 P-value=.079 n=57	Rho=−0.34 P-value=.012 n=55	Rho=−0.370 P-value=.005 n=57
Social dimension	Rho=−0.25 P-value=.072 n=55	Rho=−0.25 P-value=.074 n=53	Rho=−0.291 P-value=.031 n=55
Psychological/cognitive dimension	Rho=−0.27 P-value=.038 n=60	Rho=−0.40 P-value=.002 n=58	Rho=−0.426 P-value=.001 n=60
Energy/vitality	Rho=−0.09 P-value=.528 n=57	Rho=−0.21 P-value=.119 n=55	Rho=−0.197 P-value=.142 n=57
General health status	Rho=−0.34 P-value=.008 n=60	Rho=−0.38 P-value=.003 n=58	Rho=−0.438 P-value<.001 n=60
Total score	Rho=−0.29 P-value=.028 n=59	Rho=−0.36 P-value=.006 n=57	Rho=−0.400 P-value=.002 n=59

Spearman's correlation coefficients (Rho).

Assessment of tolerability of drug therapy for RA is a largely unexplored condition. Despite this, the treatment compliance rate is directly related to these non-serious AEs, and they are sometimes a reason for treatment withdrawal. In clinical trials, PROs are essential and show the patient's perspective in the conduct of a trial with a new drug. Similarly, in daily clinical practice, including patient's perspective in the management of their disease is on the increase, especially in chronic diseases.4,17,18

We aimed to show in the present study that the TOL-AR-18 questionnaire could present good psychometric properties, although it presents some limitations. It has a very low rate of non-response, so its administration at rheumatology visits is feasible. The high internal consistency of its items, greatly exceeding the expected theoretical values,19 is an excellent indicator of their homogeneity, all forming part of the same construct: tolerability. On the other hand, the questionnaire's low scores are consistent with the profile of the population studied, with low disease activity and little presence of AEs, it is indicating a good construct validity. In addition, we also observed a trend towards having lower tolerability (higher score) in patients with more active disease, but this is not statistically significant. This means that the TOL-AR-18's discriminant capacity would probably be confirmed were the questionnaire to be administered to a more heterogeneous population in terms of disease activity and presence of AEs. This hypothesis would need to be confirmed with further studies. Finally, correlations with the ARTS questionnaire were moderate–low, given that this questionnaire (ARTS) assesses treatment satisfaction, not directly tolerability as TOL-AR-18 does. Although the tolerability dimension (items 9 and 10 of the ARTS) theoretically assesses this concept, the TOL-AR-18 questionnaire assesses specific AEs, whereas the ARTS asks in general about side effects of treatment and maximum degree of difficulty. This degree of trouble was not very present in the study patients, as shown by the low percentage of patients who assessed the degree of difficulty due to an AE with the maximum score. Specifically, in relation to adverse events reported, after the study we were unable to obtain more powerful conclusions. Likewise, the information contained in the questionnaire was 100% reflected by the patients and that it coincides mostly with those selected by the rheumatologists. Therefore, the TOL-AR-18 is more specific for assessing tolerability.

Patient involvement in the management of their disease is becoming increasing important, as is having their self-assessment of various aspects of health available. Traditionally, tolerability in any disease is assessed from the clinical perspective of the health personnel in taking care of patient management and follow-up, by means of laboratory tests or by lists of AEs.20,21 It is also obtained by asking the patient directly. However, the patient may be asked very general questions or questions aimed at AEs that are more typical or serious or that can be expected depending on treatment regimen. Hence, the patient may not mention a very mild symptom because he/she does not believe it to be important, because he/she is not asked about it directly or because he/she believes it is normal and therefore does not report it. This study were conducted to obtain a useful, simple-to-use and reproducible tool and to validate its utility in daily clinical practice.

As a result of the need to obtain information about tolerability from the patient's perspective that may not have been obtained in the course of clinical assessment, questionnaires have been designed and validated in areas such as psychiatry, ophthalmology and gastroenterology9,22,23; but to date no questionnaires have been launched to assess the impact of AEs in the area of Rheumatology, specifically in RA.

We believe that, with the TOL-AR 18, an easy-to-use, standardized examination tool, with good measurement properties that can be self-administered by patients, available for professionals who care for RA patients on a daily basis, we answer the need that currently exists. It complements these professionals’ medical knowledge without overloading a medical visit. This questionnaire not only quantifies treatment-related symptoms, but also is valid for detecting potential adverse reactions not revealed with other kind of assessment. It also offers the novelty of quantifying the degree of trouble or importance for the patient of a particular AE, which could be greater than previously believed by the clinician. Up to now, this feature of the TOL-AR-18 might not have been gathered during visits, since it has to do with a more subjective aspect. Given that the TOL-AR-18 allows any health professional to perform a quick, detailed examination of AEs, it may be useful not only for rheumatologists, but also nursing staff and in general all professionals involved in some way in follow-up of a patient with RA.

The information obtained with this questionnaire will provide additional information about the different therapeutic strategies available and will help to optimize management of the patient with RA in terms of decision-making. Decisions might concern changes in treatment, dose or even route of administration. Thus, the information obtained will improve the patient's satisfaction with his/her treatment and HRQoL and involve him/her in management of the disease, and as a result the physician–patient relationship will be strengthened. Our results suggest a potential extrapolation to other types of rheumatic and joint diseases with therapeutic arsenals similar to those of RA. However, this will require the relevant validations of the TOL-AR-18 questionnaire in other patient populations.

This study did have some limitations. First and foremost, we consider that using DAS28 index and the number of AEs to asses the construct validity of TOL-AR-18 could be complicated, because disease activity may influence the perception of side effects by the patient. Maybe patients with high disease activity may tolerate better minor side effects than those with lower disease activity since they are more dependent of the beneficial effect of the treatment. Further studies to asses this issue may help to better understand the behaviour of the patients. Other limitations of particular note are the following: (1) Some limitations were detected in the questionnaire validation phase, such as a limited patient sample size (66 patients). Furthermore, owing to the effect of chance (consecutive enrolment), the sample turned out to be rather homogeneous with respect to disease activity and presence of AEs. Some of the questionnaire's measurement properties could probably have been assessed better with a larger, more heterogeneous sample. (2) The limitations inherent to a cross-sectional study hindered assessment of properties such as test–retest reliability, i.e. reproducibility of responses in patients with stable or similar clinical conditions or sensitivity to change that could have allowed changes in TOL-AR-18 scores to be assessed during a period of time in comparison with clinical variables. 3) The fact that a satisfaction questionnaire for patients with osteoarthritis such as the ARTS was chosen to analyze its correlation with the TOL-AR-18 might initially seem questionable. It should be noted that this questionnaire was originally designed and validated in a population with RA24 although in the Spanish validation it was used in patients with osteoarthritis. How there are not questionnaires about tolerability validated in Spanish and ARTS has a specific domain (two items) for tolerability, we chose this questionnaire instead of a generic questionnaire. Despite this, it was decided to use it to assess the effects of treatments for rheumatic and joint diseases, we believed to be more appropriate than a generic or HRQoL questionnaire for RA as it had a tolerability dimension.

To summarize, the TOL-AR-18 questionnaire is hereby presented as a support tool for professionals involved in the management of RA. It offers them a standardized means of exploring and detecting potential AEs and furthermore strengthens the physician–patient relationship by involving the patient in the disease follow-up and outcome. We therefore, due to the limitations of our study suggest, that TOL-AR-18 is a complementary tool in the management and evaluation of tolerability in RA. We suggest that the use of the questionnaire in daily clinical practice could ease a more fluent and patient-oriented management, leading to optimal control of patients with chronic arthritis.

None for the topic of the project.