Abstract
Background Assessing medication adherence in rheumatoid arthritis (RA) is clinically significant as low adherence is associated with high disease activity. Self-reported medication adherence surveys have been shown to have problems with overestimation of adherence due to social desirability bias. However, no MTX adherence studies adjusted for social desirability have been conducted to date. This study aimed to evaluate adherence to MTX and perform an investigatory search for factors associated with MTX adherence including social desirability.
Methods This cross-sectional multicenter study was conducted among adult RA patients consuming oral MTX for ≥3 months. We examined the distribution of MTX adherence, according to the eight-item Morisky Medication Adherence Scale (MMAS-8). Furthermore, an exploratory factor analysis involving social desirability was examined to identify factors associated with MTX adherence using linear regression analysis. To deal with missing values, we used multiple imputations with chained equations methods.
Results A total of 165 RA patients were enrolled. The median age was 64 years, and 86.1% were women. Based on the MMAS-8, low, medium, and high adherences were noted in 12.1%, 60.0%, and 27.9% of participants, respectively. High social desirability (coefficient, 0.14; 95% confidence interval [CI], 0.05–0.23; p<0.05) and high age (coefficient per 10 years, 0.16; 95% CI, 0.01–0.03; p<0.05) were associated with high MTX adherence, whereas full-time work was negatively associated with high MTX adherence (coefficient, -0.50; 95% CI, -0.95–-0.05; p<0.05).
Conclusions Sixty percent of RA patients had medium adherence to MTX. High social desirability, high educational level, and non-full-time work may be associated with high MTX adherence. Physicians should confirm MTX adherence before switching or adding disease-modifying anti-rheumatic drugs in cases of uncontrolled disease activity.
Background
Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized by chronic inflammation. RA affects the joints and leads to functional disability in the absence of appropriate treatment [1]. The 2015 American College of Rheumatology (ACR) and the 2019 European League against Rheumatism recommendations suggested the early use of disease-modifying antirheumatic-drugs (DMARDs), usually methotrexate (MTX), in most active RA patients [2,3]. Several studies have reported that low medication adherence is associated with high RA disease activity [4,5]. Hence, MTX adherence in RA patients is crucial.
Several studies on adherence to MTX have been conducted. However, MTX adherence differed among the studies because of the various definitions of adherence, different follow-up durations, and heterogeneity of patient populations. A previous study reported that MTX adherence is associated with age, sex, race, RA disease activity, patient beliefs about the medication, disease duration, mental health, and socio-economic status (SES) [4,6–10]. In a systematic review, beliefs on the necessity and efficacy of MTX, absence of low mood, mild disease, and MTX monotherapy were identified as the most reliable variables related to MTX adherence [11].
Social desirability is well known to influence the use of self-reported questionnaires in research. A previous study reported that social desirability is the tendency to respond to self-reported items in a way that reflects better on the respondent rather than acting accurately and truthfully [12]. Self-reported medication adherence surveys have been shown to have problems with overestimation of adherence due to social desirability bias [13,14]. A previous MTX adherence study has reported the need to control this bias [15]; however, no studies have investigated the MTX adherence survey by controlling for social desirability. Thus, this study aimed to evaluate the adherence of RA patients to MTX and exploratively evaluate factors associated with MTX adherence with the addition of social desirability.
Methods
Study design
This cross-sectional study was conducted in four hospitals: Showa University Hospital, Showa University Northern Yokohama Hospital, Showa University Fujigaoka Hospital, and Kanto Rosai Hospital.
Patient population
Participants were outpatients who met the 2010 ACR criteria for RA [16]. The inclusion criteria were age ≥20 years and consumption of oral MTX for ≥3 months. The exclusion criteria were patients with dementia, restlessness, and severe psychiatric disorders. Sequential sampling was employed. Patients were recruited between August 2013 and October 2014.
Data collection
RA disease activity was assessed using the Disease Activity Score in 28 joints (DAS28-ESR). DAS28 provides a global summative and continuous score for disease activity assessment and has been widely used in clinical trials and practice [17]. Activities of daily living (ADL) was assessed using the modified Health Assessment Questionnaire (mHAQ), which is a self-reported questionnaire that measures function, including ADL performance. Depression state was defined using the Centre for Epidemiological Studies-Depression (CES-D) scale, which includes a 20-item questionnaire; the scores range from 0 to 60. We determined that a total score of ≥16 describes a depressive state [18]. The patient’s perceptions towards medications were assessed using the Belief about Medicines Questionnaire-Specific (BMQ-specific), which includes two domains, BMQ-Specific necessity (5 items) and BMQ-Specific concern (5 items). The responses are recorded on a 5-point Likert Scale (ranging from 1=strongly disagree to 5=strongly agree). The score by domains ranges from 5 to 25; higher scores indicate stronger beliefs in the concepts represented by the scale. We used the BMQ necessity-concern differential (“BMQ-Specific necessity” minus “BMQ-Specific concern”) which shows appropriate reliability and was used in numerous studies in multivariable analysis [19]. Pain severity was assessed using the Brief Pain Inventory (BPI). Moreover, we measured the following: worst numerical rating scale (NRS) pain score, least NRS pain score, average NRS pain score, and current NRS pain score. We used the average NRS pain score other clinical studies have applied in the multivariable analysis. Social desirability is typically assessed using the Social Desirability Scale (SDS) developed by Crowne and Marlowe in 1960, with 33 items [20]. Furthermore, a 13-item SDS was developed with confirmed validity and reliability [21]. In this study, we used the 13-item SDS. Short Form-8 (SF-8) is a health-related quality of life (QoL) questionnaire, which measures the same eight health domains (physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health) similar to Short Form-36 (SF-36v2®) but shorter, using eight questions and whose validity and reliability were confirmed [22]. This questionnaire was used to assess the QoL in RA [23]. The score range of each domain is from 0 to 100; higher scores suggest better health-related QoL. We used two summary scores: mental component summary (MCS; showing mental status) and physical component summary (PCS; showing physical status). The Japanese versions of mHAQ, CES-D, BMQ, BPI, SDS, and SF-8 were validated [24,25].
Additionally, we obtained data on SES, including marital status, educational level, employment status, and living status. Data on patient characteristics, medication, and SES were further collected using questionnaires. Other data were obtained from medical chart records. After consulting the doctor, the participants answered the Morisky Medication Adherence Scale (MMAS-8), mHAQ, CES-D, BMQ, BPI, SDS, SF-8, and SES questionnaires; they returned the completed questionnaires to the data center by mail.
Outcome
The primary outcome was the distribution of MTX adherence according to MMAS-8 at the time of the survey based on the patient’s description. MMAS-8 consists of an eight-item questionnaire [26–28]. MMAS-8 scores range from 0 to 8 and are classified as follows: low, <6; medium, >6 and <8; and high, 8. Originally, MMAS-8 was developed for daily administration of oral medicine. MTX was administered once a week; thus, we developed the Japanese questionnaire for MTX accordingly. First, we translated the questionnaire from English to Japanese and resolved some language discrepancies. Second, back translations were performed. Third, we discussed the questionnaire and achieved a consensus. Finally, a pre-test in five RA patients followed by necessary modifications were performed to finalize the Japanese questionnaire.
Statistical analyses
MMAS-8, sex, CES-D, and SES were used as categorical variables, whereas age, MTX dose, MTX dosing frequency, duration of MTX treatment, disease duration, DAS28-ESR, mHAQ, BMQ, BPI, SDS, and SF-8 were used as continuous variables. Summary statistics were presented as median with interquartile range (IQR) and numbers with proportion (%). First, we evaluated the distribution of MTX adherence according to MMAS-8. Subsequently, we compared MMAS-8 and pre-described factors using one-way analysis of variance (ANOVA), Kruskal-Wallis test, and chi-squared test. For significant factors, we conducted a trend analysis. Finally, multiple linear regression analysis was performed to exploratively assess factors associated with MTX adherence with the addition of social desirability. The co-variables selected were as follows: age, sex, disease duration, RA disease activity (DAS28-ESR), depression state (CES-D), reliability of the medication (BMQ necessity-concern differential), social desirability (SDS), educational level (more than, equal to college-level, or not), and employment status (full-time work or not). These factors were further selected based on previous studies and clinical importance [4,6–10,29–31].
This was an observational study; hence, the sample size was not calculated on a statistical basis but was measured to the extent practicable over the study period. To compare the mean DAS-28 scores among the three groups using ANOVA, the effect size (small, 0.1; medium, 0.25; and large: 0.4) customarily proposed by Cohen [32] was set at medium (0.25), the significance level was set at 5% on both sides, and the power was set at 80%. The total number of cases was calculated to be 159. The target number of cases was 176, assuming a dropout rate of 10%.
To deal with missing values, we used multivariable multiple imputations with chained equations methods to increase power and minimize selection bias. We included age, sex, disease duration, RA disease activity, depression state, reliability of the medication, social desirability, educational level, and employment status for each imputation model. We generated 10 imputed datasets and combined coefficient estimates using Rubin’s rules for each imputation. Imputation of missing data in this way assumes that data were missing at random since any reasons for missing were explained by the observed variables included in the imputation model.
A two-sided p-value <0.05 was considered statistically significant. All statistical analyses were conducted using STATA 14.2 (StataCorp LP, College Station, TX) software.
Ethical considerations
The ethics committee of Showa University Hospital (approval number 1446) and Showa University Toyosu Hospital, Showa University Northern Yokohama Hospital, Kanto Rosai Hospital approved this study, and informed consent was obtained from all participants before study enrolment. All study procedures were performed in accordance with the Declaration of Helsinki. Patient information was anonymized and de-identified before analysis.
Results
Patient flow chart
Figure 1 shows the patient flow chart. Initially, 181 RA patients were invited. Of 169 patients eligible for the study, four were excluded because of missing MTX adherence data (n=2) and clinical data (n=2). A total of 165 RA patients were included in the final analysis.
Baseline characteristics
The characteristics of the participants are shown in Table 1. The median age of the patients was 64 years (IQR, 54–72), and 86.1% were women. The median dose of MTX was 8 mg per week (IQR, 6–12), median MTX dosing frequency was three times per week (IQR, 2–3), and median duration of MTX treatment was 36 months (IQR, 17–75). The proportion of missing data varied from 0 to 15.2% (Supplementary table 1).
Distribution of MTX adherence
Based on MMAS-8, low, medium, and high adherence rates were noted in 12% (20/165), 60% (99/165), and 28% (46/165) of cases, respectively. The median MMAS-8 score was 7 (IQR, 6.5–8).
Patient’s characteristics by grade of MTX adherence
Patient’s characteristics by grade of MTX adherence divided by MMAS-8 are shown in Table 1. A significant association between levels of MTX adherence and age, mHAQ, SDS, SF-8 (PCS), and employment status was found. Based on trend analysis, older participants had significantly high adherence, and the higher the social desirability, the higher the adherence. However, a significant association between MTX adherence and mHAQ and SF-8 (PCS) was not observed in the trend analysis.
Factors associated with high MTX adherence considering social desirability
In the multiple linear regression analysis adjusted for age, sex, disease duration, RA disease activity (DAS28-ESR), depression state (CES-D), reliability of the medication (BMQ necessity-concern differential), social desirability (SDS), educational level (more than or equal to college-level or not), and employment status (full-time work or not), higher social desirability (coefficient, 0.14; 95% confidence interval [CI], 0.05–0.23; p<0.05) and higher age (coefficient per 10 years, 0.16; 95% CI, 0.01–0.03; p<0.05) were associated with high MTX adherence, whereas full-time work was negatively related to high MTX adherence (coefficient, -0.50; 95% CI, -0.95–-0.05; p<0.05) (Table 2).
Discussion
Elderly participants and high social desirability were significantly associated with high adherence based on trend analysis. Multiple linear regression analysis showed that high MTX adherence in RA patients is associated with high social desirability, high educational level, and non-full-time work.
In our study, only 27.9% of the participants showed high medication adherence. Several studies reported that adherence to DMARDS was 40%–107% [9,33,34]. There are several potential reasons for the observed low adherence. First, the method of adherence measurement in each study was different. There were interviews, questionnaires, tablet count, drug concentration, and doctor’s judgment as measurements. A previous study showed that the adherence to interview, tablet count, drug concentration, and doctor’s judgment were 96%, 77%, 58%, and 42%, respectively, which showed vast differences [35]. Since each measurement has advantages and disadvantages, the gold standard is not determined. We chose a self-reported questionnaire in this study since this scale is not expensive or invasive and can be possibly validated. Second, the time point of adherence measurement was different between the studies. We measured adherence during the maintenance period (mean disease duration, 53 months). A previous report suggested that the adherence of DMARDs in the early period was low (58%) [5]. Another study reported the adherence to salazosulfapyridine as 87% at three years from the start of medication [30]. In contrast, the adherence to DMARDs in the maintenance period varied from 84%,4 70% [36], to 95.4% [37]. Previous reports have suggested that the adherence to DMARDs was not different between the early and maintenance periods. It is crucial to note that the time point of measuring medication adherence is the maintenance period because RA is a chronic disease that requires DMARDs for a long duration. The third is the difference in participant characteristics. Living status and marital status affected adherence [10,38]. Our study could not adjust the living status and marital status because we could not include more factors based on our sample size. In addition, the age of the participants may have had an effect. In the original MMAS-8 article [26], the mean age was 52 years; however, in our study, the median age was 64 years, which means that the participants in our study were older, and this may have influenced the results. Furthermore, older people may not have revealed the truth. Thus, when referring to drug adherence studies, the method of measuring adherence, whether the target patient is in the disease’s early or maintenance phase, and the patient’s characteristics should be confirmed before adaptation.
Our study detected high social desirability, high educational level, and non-full-time work as factors affecting MTX adherence. A previous study reporting MTX adherence found that adherence is associated with age, sex, race, RA disease activity, patient beliefs about the medication, disease duration, mental health, and SES. The discrepancy between our findings and those of other studies could be attributed to three reasons. First, we measured and adjusted social desirability, which possibly affected the answers in the questionnaire. Although a recent MTX adherence research mentioned the necessity to account for this bias [15], no studies have examined how MTX adherence survey accounts for social desirability. The results may have differed since we conducted a factor search that included social desirability. Second, racial differences may have led to different outcomes. Our study was based on a Japanese population; however, previous studies were mainly on white people, black people, and Hispanic people. Canon et al. reported that the adherence to MTX was associated with the Caucasian race [4]. Therefore, adherence may be influenced by race. Third, the definition of medication adherence varies between studies. We used MMAS-8 in our study. Conversely, previous studies used other methods to assess medication adherence, such as physicians’ estimation, different self-reported scales, drug concentration, tablet count electronic monitors, and other questionnaires. A prior study reported that medication adherence differs because the measurement method employed seems to vary [11]. Therefore, we need to practice caution when using research results because the results may differ depending on the method of measuring adherence and the variables incorporated, including social desirability.
Our study demonstrated that RA patients with high social desirability had high MTX adherence. A previous study reported that social desirability refers to responding to self-reported items that make the respondent seem better [12]. Contrastingly, the influence of social desirability is unknown, although its significance is generally accepted. A previous report has suggested that social desirability does not play a major role in assessing subjective well-being [39]. Therefore, although the effect of social desirability on drug adherence may not be significant, we were able to search for factors that consider social desirability.
Our study had three strengths. First, we assessed social desirability, leading to social desirability bias. To our knowledge, no clinical studies adjusting social desirability have investigated the association between MTX adherence and various factors. Second, to our knowledge, this is the first research to examine MTX adherence in an Asian population. Third, the response rate to the questionnaires was high (93.4%), thereby strengthening internal validity.
This study has significant implications for the clinician. Our findings may prevent unnecessary DMARDs changes since physicians tend to overestimate patients’ medication adherence [40]. They should confirm MTX adherence before switching or adding DMARDS. Accordingly, this provides an opportunity to reduce the healthcare cost and adverse events of medication.
However, this study had several limitations. First, in the multivariable analysis, some immeasurable essential factors (such as characteristics of each participant, joint deformities, number of medications, and MTX dosing frequency) that could affect MTX adherence possibly exist. We did not determine patient personality due to the difficulty in quantifying it. In addition, we did not collect radiographs even if they are necessary for determining joint deformity because they are burdensome for patients. However, we adjusted for joint deformity using the mHAQ to reduce its possible influence on the findings. In addition, although the number of MTX doses was collected, it was not possible to incorporate variables due to sample size limitations. Second, the generalizability of our findings is limited. Our study setting was primarily university hospitals in Japan; thus, the results of this study cannot be applied to the clinical setting, in small- and medium-sized hospitals, or in settings including other races. Furthermore, the highest dose of MTX in Japan in the study was 8 mg per week. This dose is less than the standard dose of 20 mg worldwide, which limits the indications. Third, we used a self-reported questionnaire of adherence. Hence, the completed questionnaire possibly influenced social desirability. Although we could not adjust this bias completely, social desirability was added as a factor for analysis.
Conclusions
Our results demonstrated that sixty percent of the RA patients had moderate adherence to MTX. Moreover, high MTX adherence in RA patients is associated with high social desirability, high educational level, and non-full-time work.
Data Availability
The dataset analyzed in this paper is available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The ethics committee of Showa University Hospital (approval number 1446) and other each institution approved this study, and informed consent was obtained from all participants before study enrolment. All study procedures were performed in accordance with the Declaration of Helsinki. Patient information was anonymized and de-identified before analysis.
Consent for publication
Not applicable.
Availability of data and materials
The dataset analyzed in this paper is available from the corresponding author on reasonable request.
About MMAS8
The use of the MMAS diagnostic adherence assessment instrument is protected by US copyrighted and trademarked laws. Permission for use is required. A license is available from MORISKY MEDICATION ADHERENCE RESEARCH, LLC., Donald E. Morisky, ScD, ScM, MSPH, MMAR, LLC, 294 Lindura Ct., Las Vegas, NV 89138; dmorisky{at}gmail.com.
Competing interests
Donald E. Morisky holds a copyright for the Morisky Medication Adherence Scale-8, is one of its authors, and collects fees in exchange for licenses to use the scale in research.
This does not alter the authors’ commitment to objectivity in research or adherence to data sharing policies. He was not involved in any data analysis. For the remaining authors, none were declared.
Funding
The authors have not received funding to conduct this study.
Authors’ contributions
Nobuyuki Yajima: Methodology, Project administration, Writing – original draft, Writing – Review & Editing
Takashi Kawaguchi: Supervision, Conceptualization, Project Administration, Writing – original draft, Writing – Review & Editing
Ryo Takahashi: Data curation, Writing – original draft, Writing – Review & Editing
Hiroki Nishiwaki: Data curation, Writing – original draft, Writing – Review & Editing
Youichi Toyoshima: Data curation, Writing – original draft, Writing – Review & Editing
Koei O: Data curation, Writing – original draft, Writing – Review & Editing
Tsuyoshi Odai: Data curation, Writing – original draft, Writing – Review & Editing
Takayuki Kanai: Data curation, Writing – original draft, Writing – Review & Editing
Donald E. Morisky: Supervision, Writing – original draft, Writing – Review & Editing
Takuhiro Yamaguchi: Formal Analysis, Supervision, Writing – original draft, Writing – Review & Editing
Tsuyoshi Kasama: Supervision, Writing – original draft, Writing – Review & Editing
Acknowledgments
We thank Noboru Murata (Kikuna Memorial Hospital) and Osamu Namiki for critically reviewing the manuscript.
List of Abbreviations
- ADL
- Activities of Daily Living
- BMQ
- Beliefs about Medicines Questionnaire
- BPI
- Brief Pain Inventory
- CES-D
- Center for Epidemiological Studies Depression
- DAS28-ESR
- Disease Activity Score (28 joint count)-erythrocyte sedimentation rate
- DMARDs
- disease-modifying antirheumatic-drugs
- MCS
- mental component summary
- mHAQ
- modified Health Assessment Questionnaire
- MMAS-8
- Morisky Medication Adherence Scale MTX: methotrexate
- NRS
- Numerical Rating Scale
- PCS
- physical component summary
- RA
- rheumatoid arthritis
- SDS
- Social Desirability Scale
- SF-8
- 8-item Short-Form Health Survey
References
