1. Top of page
  2. Abstract
  7. Acknowledgements


Patients and physicians often differ in their perceptions of rheumatoid arthritis (RA) disease activity, as quantified by the patient's global assessment (PGA) and by the evaluator's global assessment (EGA). The purpose of this study was to explore the extent and reasons for this discordance.


We identified variance components for the PGA and EGA in RA patients who were starting therapy with methotrexate in an academic outpatient setting. We analyzed predictors of the observed discrepancy in these measures (calculated as the PGA minus the EGA) and in their changes (calculated as the PGAchange minus the EGAchange).


We identified 646 RA patients, and among them, 77.4% of the variability in the PGA and 66.7% of the variability in the EGA were explainable. The main determinants for the PGA were pain (75.6%), function (1.3%, by Health Assessment Questionnaire), and number of swollen joints (0.5%); those for the EGA were the number of swollen joints (60.9%), pain (4.5%), function (0.6%), C-reactive protein (0.4%), and the number of tender joints (0.3%). Increased pain led to a discrepancy toward worse patient perception, while increased numbers of swollen joints led to a discrepancy toward worse evaluator perception, both explaining 65% of the discordance between the PGA and the EGA. Likewise, changes in pain scores and numbers of swollen joints proved to be the main determinants for discrepant perceptions of changes in RA disease activity, explaining 34.6% and 12.5% of the discordance, respectively.


The most significant determinants for the cross-sectional and longitudinal discrepancy between the PGA and the EGA are pain and joint swelling, respectively. Understanding the reasons for a discordant view of disease activity will help to facilitate the sharing of decision-making in the management of RA.

Rheumatoid arthritis (RA) is a chronic inflammatory musculoskeletal disease characterized by pain, stiffness, swelling, and tenderness of the synovial joints, ultimately leading to joint destruction, disability, and reduced quality of life (1–5). The goal of RA treatment, therefore, is to prevent these sequelae by interfering with the inflammatory process. This goal was recently formalized by an international task force (6). An important aspect of the management of RA is a thorough evaluation and discussion of treatment decisions between the patients and their physicians, often referred to as “shared decision-making” (6, 7).

However, the patients' opinions do not always match those of their physicians: a discrepancy between patients' and physicians' ratings of physical functioning and general health has been demonstrated (8, 9). The direction of the discrepancy usually points toward a better rating by the physicians than by the patients themselves, while physicians are more concerned about the potential risks associated with the treatments used for achieving a good disease activity outcome than the patients are (8, 9). The consequence of such a discordant viewpoint with regard to disease activity is that decisions are often prone to not being shared between patients and physicians. A frequently used means of quantifying the overall perception of disease activity is through an estimation of global disease activity, both by the patient (patient's global assessment [PGA]) and by the physician (evaluator's global assessment [EGA]). These two assessments are part of the core sets of RA disease activity measures that were defined in the early 1990s (10–12).

The quest for understanding the reasons for discrepancies in evaluations of disease activity becomes particularly important in the context of recent recommendations that define remission as the treatment target (6), especially considering that the new Boolean-based American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) remission criteria require the PGA to be ≤1 cm (13). It has already been reported that a disproportionately high PGA may prevent many patients from being classifiable as in remission (14). Furthermore, it has been suggested that composite measures consisting solely of patient-reported outcomes be used for disease activity assessment, rather than the more-traditional composite scores that include the EGA and/or joint counts (15, 16). Since treatment decisions are based on such scores, a significantly discrepant judgment about disease activity rendered by patients and their physicians in either direction may have untoward therapeutic consequences (17).

Taking all of these concerns into consideration, our aim in the present study was to investigate the factors that underlie the observed discrepancy between the patient's and the physician's judgments of disease activity. For this purpose, we evaluated the 2 global measures, PGA and EGA, in a variety of cross-sectional and longitudinal analyses.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Patients and data.

We identified study patients from a large observational RA patient database, where clinical and laboratory variables characteristic of RA are assessed routinely every 3–4 months and are prospectively documented. The quality of the data is ensured by constant maintenance and periodic quality checks (18–20). Patients enrolled in this database are diagnosed as having RA according to the ACR 1987 revised criteria (21) or the ACR/EULAR 2010 classification criteria (22), depending on how long patients have been in our care. These patients are seen regularly and are treated by rheumatologists or rheumatologists in training. Data from the baseline visit during which methotrexate (MTX) therapy was initiated at our clinic and from 1 subsequent visit (followup visit), which occurred 2–7 months later (average of 4 months), were identified.

Information on demographic features and on all core set measures were extracted from the database for both time points. These data included age, sex, disease duration (defined as time since diagnosis), numbers of swollen and tender joint counts, presence of rheumatoid factor (RF) and/or anti–citrullinated protein antibodies (ACPAs), the C-reactive protein level (CRP), erythrocyte sedimentation rate (ESR), scores on the Health Assessment Questionnaire (HAQ) disability index (23), duration of morning stiffness, patient's assessment of pain, as measured using a 100-mm visual analog scale (VAS), and global assessments of disease activity by the evaluator/physician (EGA) and the patient (PGA). The EGA was determined without knowledge of the patient's laboratory test results. The question for determining the PGA was, “How do you estimate your disease activity today” (in German, “Wie schätzen Sie heute Ihre Krankheitsaktivität ein?”). The anchors for the VAS are defined as “no disease activity” (0 mm) and as “highly active disease” (100 mm). The question for determining the patient's assessment of pain was, “How severe is your pain today” (in German, “Wie stark sind Ihre Schmerzen heute?”). Only patients with complete data on all core set measures were used in our analyses (24).

Analyses of the determinants of the PGA and the EGA.

In order to identify variables correlating with the PGA and the EGA, we calculated Pearson's correlation coefficients. The variables tested included core set measures as well as morning stiffness and RF. In the next step, the significant variables identified from this univariate correlation were tested in multivariate analyses for between-subject effects and were further tested by linear stepwise multivariate regression modeling to determine their independent contribution to the PGA and the EGA. In these models, variables were selected using a cut point of P < 0.05 for entry into the models and P < 0.1 for removal from the models. These analyses were done separately for each of the baseline and followup visits to see whether the results were stable over time.

Analyses of determinants of discrepancies between the PGA and the EGA.

To analyze discrepancies between the PGA and the EGA, we calculated the new variable “PGA minus EGA” (PGA – EGA); the value of this variable quantifies the degree of discrepancy between the two measures. Therefore, the range of this variable is from –100 mm to +100 mm, with negative values indicating worse estimation of disease activity by the physician and positive values worse estimation of disease activity by the patient. We then again used Pearson's correlational analyses to identify variables that were to be tested in a multivariate stepwise regression model. For purposes of graphic illustration, we calculated estimated marginal means of the PGA – EGA. These analyses were performed with data from all baseline visits.

Analyses of determinants of discrepancies in improvement between the PGA and the EGA.

The steps described above were repeated for the longitudinal analysis by investigating the consistency of changes in perception over time by the patients and the physicians. For this purpose, we calculated change scores for disease activity variables based on the respective values at baseline and followup. The newly created variable “PGAchange minus EGAchange” (PGAchange – EGAchange) theoretically ranges from –200 mm to +200 mm, and can be interpreted as representing the difference in the perception of longitudinal changes between the patient and the physician. We divided patients into 3 groups according to the PGAchange and into 3 groups according to the EGAchange. Both groupings represented worsening of disease activity (increase of >5 mm), unchanged disease activity (within –5 mm to +5 mm from baseline), and improvement in disease activity (decrease by >5 mm). These groups were then cross-tabulated separately for PGAchange and EGAchange. For assessment of differences between patient groups with improvement and groups with worsening of the core set of variables, analysis of variance and analysis of covariance were performed.

To identify determinants of the discordance between change in the PGA and change in the EGA, we first correlated baseline PGA and EGA values, as well as change scores for the core set of variables with the PGAchange – EGAchange (Pearson's r), and then subjected the significant variables to a stepwise multivariate linear regression model, similar to the analyses in the initial cross-sectional population.

All analyses were performed using SPSS software version 17.0.


  1. Top of page
  2. Abstract
  7. Acknowledgements


In total, we identified 646 patients in whom MTX was initiated, and data for all of these patients were used in the cross-sectional analyses. Of these 646 patients, 437 attended a followup visit within 2–7 months (mean 4 months), and data from the baseline and followup visits were used in the longitudinal analysis. Demographic characteristics of the study patients at the baseline and followup visits are shown in Table 1. Most patients were female (80%), positive for ACPA (80%) or RF (68%), and on average, showed moderate disease activity. The mean duration of RA at baseline was 7.7 years, which is partly due to the fact that 30% of the study patients had longstanding RA (≥10 years), and many patients have already been diagnosed as having RA when they first present to our clinic (in 27% of patients, the diagnosis was established during the baseline visit).

Table 1. Demographic and clinical characteristics of the patients at initiation of MTX and at followup*
 BaselineFollowup (n = 437)
All patients starting MTX (n = 646)Only patients with followup (n = 437)
  • *

    Data obtained at baseline, when methotrexate (MTX) was initiated at our clinic, are subcategorized according to those who did and those who did not return for at least 1 followup visit 2–7 months later (mean 4 months). CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; VAS = visual analog scale (0–100 mm); PGA = patient's global assessment; EGA = evaluator's global assessment; HAQ = Health Assessment Questionnaire; ACPA = anti–citrullinated protein antibody; RF = rheumatoid factor; SDAI = Simplified Disease Activity Index; CDAI = Clinical Disease Activity Index; DAS28 = Disease Activity Score in 28 joints.

Age, mean ± SD years56 ± 1456 ± 14
% female8081
Disease duration, mean ± SD years7.7 ± 9.97.2 ± 9.8
CRP, mean ± SD mg/dl1.6 ± 2.51.5 ± 2.31.2 ± 1.9
ESR, mean ± SD mm/hour30 ± 2430 ± 2426 ± 20
Pain score, by VAS, mean ± SD mm36 ± 2634 ± 2430 ± 24
Global assessments, mean ± SD mm   
 PGA39 ± 2738 ± 2533 ± 25
 EGA23 ± 2123 ± 2118 ± 18
 PGA minus EGA16 ± 2515 ± 2315 ± 24
Joint counts, mean ± SD (28 assessed)   
 Swollen joints3.7 ± 4.53.4 ± 4.32.5 ± 3.4
 Tender joints4.4 ± 6.14.1 ± 5.73.0 ± 4.9
Morning stiffness, mean ± SD minutes40 ± 7336 ± 6826 ± 55
HAQ score, mean ± SD0.9 ± 0.80.8 ± 0.80.7 ± 0.7
% ACPA positive808179
% RF positive686665
SDAI, mean ± SD16 ± 1315 ± 1211 ± 10
CDAI, mean ± SD14 ± 1214 ± 1111 ± 9
DAS28, mean ± SD4.0 ± 1.43.9 ± 1.43.5 ± 1.3

Determinants of PGA and EGA.

As shown in Table 2, among all of the variables we tested, pain scores showed an outstandingly high correlation with the PGA (r = 0.86, P < 0.001). In contrast, the swollen joint count had the strongest correlation with the EGA (r = 0.77, P < 0.001). All other variables showed low (r < 0.3) to moderate (0.3 < r < 0.5) correlations. Correlations were similar when baseline or followup data were used (Table 2).

Table 2. Correlation of demographic and disease activity variables with the patient's and the evaluator's perceptions of disease activity*
 Cross-sectional correlations (r)
  • *

    Data represent Pearson's correlation coefficients for the patient's global assessment (PGA) and the evaluator's global assessment (EGA) in 646 patients at baseline and 437 patients at followup. HAQ = Health Assessment Questionnaire; VAS = visual analog scale; CRP = C-reactive protein; RF = rheumatoid factor; ACPA = anti–citrullinated protein antibody; CDAI = Clinical Disease Activity Index.

  • P < 0.05.

  • P < 0.001.

Female sex−0.03−0.02−0.01−0.02−0.02−0.00
Disease duration−0.07−0.04−−0.06
Swollen joint count0.240.160.770.70−0.39−0.35
Tender joint count0.460.470.410.420.150.19
Morning stiffness0.400.380.
HAQ score0.490.520.300.320.260.31
Pain score by VAS0.860.870.370.370.620.64

When considering a multivariate inclusion model (i.e., including all tested variables) using the baseline data, 93% of the variation in the PGA was explained (mainly by pain, swollen joint count, and HAQ scores) and 86% of the variation in the EGA (mainly by swollen joint count, pain score, CRP level, and tender joint count). In a subsequent stepwise multivariate regression model, all variables significant at P < 0.001 in the univariate correlation (Table 2) were considered to be predictors except for the Clinical Disease Activity Index (CDAI) (25), which is a composite measure of several of them. In this model, the pain score, HAQ score, and swollen joint count remained significant determinants of the PGA, with pain explaining 75.6% of the variability, followed by only minor contributions by the HAQ score and the swollen joint count (1.3% and 0.5%, respectively), which together explained a total of 77.4% of the variability. When pain was excluded as predictor, the HAQ score (27%), tender joint count (7%), duration of morning stiffness (3%), and swollen joint count (1%) were significant in explaining a total variability of 38% of the PGA.

In the multivariate analysis, pain explained most of the association between the tender joint count and the PGA that had been seen in the univariate analysis. For the EGA, the contributions of the above-mentioned variables were quite different from their contributions to the PGA, explaining 4.5% (pain score), 0.6% (HAQ score), and 60.9% (swollen joint count) of the EGA. The CRP level and tender joint count contributed 0.4% and 0.3%, respectively, to the explanation (66.7% of the total EGA variability explained).

The exact composition of the variables contributing to the PGA and the EGA is displayed in Figure 1. At the followup visit, the results were very similar, with a total explained variability in the PGA and the EGA amounting to 76.9% and 62.1%, respectively (details not shown).

thumbnail image

Figure 1. Components of the patient's global assessment (PGA) and the evaluator's global assessment (EGA). Values of the components are the percentage of ΔR2 computed in the multivariate regression model. Values for the PGA are 22.6% unexplained, 0% C-reactive protein (CRP), 1.3% Health Assessment Questionnaire (HAQ), 0% tender joint count (TJC), 0.5% swollen joint count (SJC), and 75.6% pain. Values for the EGA are 33.3% unexplained, 0.4% CRP, 0.6% HAQ, 0.3% tender joint count, 60.9% swollen joint count, and 4.5% pain.

Download figure to PowerPoint

Determinants of discrepancies in the PGA and the EGA.

Evaluation of disease activity yielded discordant scores between the patients and their physicians in 76% of our cohort: 394 patients (61%) had higher PGA scores, 97 patients (15%) had higher EGA scores, and 155 patients (24%) had concordant PGA and EGA scores (margin of ±5 mm). Univariate analyses of all baseline visits showed that the tender joint count (r = 0.15), the duration of morning stiffness (r = 0.21), the HAQ score (r = 0.26), the swollen joint count (r = –0.39), and the pain score (r = 0.62) were highly significantly associated (P < 0.001 for each variable) with the degree of discrepancy between the PGA and the EGA (Table 2). The swollen joint count was the only variable showing a significant inverse correlation, which indicates that a higher number of swollen joints led to a worse perception by the evaluator than by the patient, whereas for the other variables, the correlations were positive, indicating a worse perception by the patient than by the evaluator.

Multivariate analyses of baseline data indicated that only the pain score and the swollen joint count were significantly and independently associated with the PGA minus EGA (P < 0.001 for each variable). The pain score (ΔR2 = 0.368) and the swollen joint count (ΔR2 = 0.285) explained 65% of the variability in the PGA – EGA, with the regression coefficients indicating opposing effects of these two variables (Table 3). However, 35% of the variability remained unexplained by this model.

Table 3. Summary of the results of regression analyses, with regression coefficients for the predictor variables*
 bSEP95% CI
  • *

    Explanatory models for the discrepancy between the patient's global assessment (PGA) and the evaluator's global assessment (EGA), as well as for the discrepancy between PGA changes and EGA changes, are shown. 95% CI = 95% confidence interval; CDAI = Clinical Disease Activity Index.

Discrepancy between PGA and EGA    
 Constant2.061.130.069−0.16, 4.28
 Pain0.690.030.0000.64, 0.73
 Swollen joint count−2.980.140.000−3.26, −2.70
Discrepancy between changes in PGA and EGA    
 Constant−3.061.530.047−6.07, −0.04
 Pain change0.360.05<0.0010.27, 0.45
 Swollen joint count change−5.400.57<0.000−6.52, −4.28
 Tender joint count change−2.470.46<0.001−3.37, −1.57
 PGA baseline0.190.04<0.0010.11, 0.26
 EGA baseline−0.300.05<0.001−0.39, −0.20
 CDAI change2.620.38<0.0011.86, 3.37

Figure 2 illustrates this influence with estimated means for the PGA – EGA. When assuming an average and fixed swollen joint count (3.2 swollen joints in our patients), a pain score >12 mm produced a discrepancy toward higher PGA values. More generally, with each millimeter increase in the pain score on the VAS, the discrepancy between the two global assessments increased by 0.7 mm toward higher PGA values. In contrast, each additional swollen joint led to a 3-mm increase in the discordance toward higher EGA values. At an average and fixed level of pain (33.9 mm in our patients), concordance between the PGA and the EGA is attained at 10 swollen joints (Figure 2). A similar analysis using the data from the followup visit revealed similar findings (data not shown).

thumbnail image

Figure 2. Effects of swollen joints and pain scores on the discrepancy between the patient's global assessment (PGA) and the evaluator's global assessment (EGA) by general linear modeling (PGA – EGA). Shown are the effects of increasing pain scores (left) and increasing numbers of swollen joints (right) on discrepancies between the PGE and the EGA. Estimations were done for a cohort mean pain score of 34 mm and a cohort mean swollen joint count of 3.2, respectively. Diagonal regression lines are for the prediction of the PGA – EGA. Horizontal lines show the marginal means of the PGA − EGA. Values are the mean ± 95% confidence interval.

Download figure to PowerPoint

Determinants of discrepancies in changes in the PGA and the EGA.

Figure 3 shows the proportions of changes in PGA and EGA, respectively, and the degree of concordance in the perception of improved, worsened, or unchanged disease activity. By these definitions, 42% of patients thought that their disease activity had improved since initiation of MTX therapy and 45% of evaluators thought that their patients' disease activity had improved. A total of 37% of the patients declared that there had been no change in disease activity, which was also the opinion of 27% of the evaluators. In 21% of patients, there was a worse PGA value than at baseline, and in 28% of patients, their EGA was rated worse than at baseline.

thumbnail image

Figure 3. Frequency of concordant and discordant evaluations of changes in the patient's global assessment (PGA) and the evaluator's global assessment (EGA). Values that varied at ±5 mm were designated as unchanged. Shown are the PGA response status by groups of EGA (top) and the EGA response status by groups of PGA (bottom). Percentages within the bars indicate the agreement between the respective response categories.

Download figure to PowerPoint

Chi-square analyses showed a significant difference between changes in the PGA and EGA (P < 0.001 across groups). The greatest concordance was seen for improvement, with the EGA showing 70% concordance with the PGA and the PGA showing 63% concordance with the EGA. Analyses of variance showed that these patients with improvement, as compared to the other patients, had the highest baseline PGA and EGA values and the largest changes in pain scores, swollen joint counts, and tender joint counts. In contrast, in terms of worsening, the concordance between patients and physicians was lower, with the EGA showing 48% concordance with the PGA and the PGA showing 37% concordance with the EGA.

In univariate correlational analyses, significant associations with PGAchange – EGAchange values were observed for the baseline PGA value (r = 0.36), as well as for changes in the duration of morning stiffness (r = 0.15), the swollen joint count (r = –0.15), tender joint count (r = 0.13), HAQ score (r = 0.27), pain score (r = 0.58), and the CDAI (r = 0.19) (P < 0.01 for each variable).

Results of a stepwise multivariate regression model showed that the change in the pain score (ΔR2 = 0.346), the swollen joint count (ΔR2 = 0.125), and the tender joint count (ΔR2 = 0.036) were the main determinants of discrepancies in changes in the PGA and the EGA. Minor, but still significant effects were seen for the baseline PGA (ΔR2 = 0.026) and EGA (ΔR2 = 0.028) values, as well as for changes in the CDAI (which comprises the swollen joint count, tender joint count, PGA, and EGA; ΔR2 = 0.026). These variables remained independently associated with this discrepancy and explained 59% of the variability in the PGAchange – EGAchange values (P < 0.001 for each variable). According to this regression model, 1 additional improved swollen joint or tender joint, or 1 additional millimeter of baseline EGA increased the discordance toward higher evaluator-assessed improvement values by 5.4 mm, 2.5 mm, or 0.3 mm, respectively (Table 3). In contrast, each additional millimeter of baseline PGA and each millimeter of improved pain increased the discordance toward higher patient-perceived improvement by 0.19 mm and 0.36 mm, respectively. Greater CDAI changes also increased the discordance toward higher patient perceived improvement (2.62 mm per unit of CDAI improvement) (Table 3).

Analyses of cross-tabulated patient groups supported the finding of our regression model. Patients with improving PGA values had the greatest improvement in pain values (P < 0.001), and patients with improved EGA values showed the highest improvement in swollen joint counts (P < 0.0001). Patients with worsened disease activity by the EGA had more swelling than at baseline. However, patients who rated their disease activity as worse than at baseline had increased pain scores.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Our study clearly shows that pain is the single most important determinant of the patient's perception of RA disease activity, while for the physician/evaluator, it is mostly the joint swelling. Importantly, for both the patients and their physicians, ∼25% of the variability in the perception of disease activity cannot be explained by the core set instruments. Indirectly, this implies that pain could objectively be regarded as underappreciated by the caregivers and a source of potential disagreement between patients and their physicians when it comes to treatment decisions.

In their rating of pain, patients likely also include tenderness of the joints. This is an explanation for what, at first glance, is the surprising fact that the number of tender joints was more important to the physician than to the patient. The same argument also partly applies to the small contribution of functional disability to the patient's rating. When pain was omitted from the stepwise analysis, the HAQ score appeared as the most important predictor, consistent with its known association with disease activity (3, 26).

When speaking about the patient's global assessment, it is necessary to consider the detailed phrase of the question. In our clinic, the patients are specifically asked to assess their disease activity during the past week and not “all the ways their arthritis affects” them (10, 13, 15) or their general health. While there appears to be only little difference between these questions when using them to calculate the Disease Activity Score in 28 joints, the direct correlation between assessments of general health or disease activity as questions for the PGA was found to be only r = 0.5 (27), again emphasizing the importance of the phrasing.

While the principal importance of pain to the patient (28, 29) and swelling to the physician (30) are known, the most important aspect of our study was the identification of determinants of the discrepancy between the PGA and the EGA (31–33) and their interplay. Also, the majority of our study patients rated their disease activity worse than did their physicians. The multivariate analysis demonstrated that for the “average” patient with 3 swollen joints, the patient rating was worse than the physician rating, once the pain score exceeded a very low level of only 15 mm. In contrast, for an average pain score of 34 mm, the evaluator rating was worse than the patient rating as long as the number of swollen joints was >10. Thus, slightly elevated pain scores can shift perceptions toward discordance, and a relatively high number of swollen joints is necessary to compensate for this.

This is essentially also true for the perception of improvement in disease activity. Patients tend to weight their pain into their disease activity evaluation to an even greater extent than their physicians do for the number of swollen joints. Interestingly, patients and physicians have a higher level of agreement about the presence of improvement than about the presence of worsening. Decreases in joint activity always lead to an improved rating of disease activity by the physician, but this is generally not so in the perception of patients, particularly those whose pain levels had increased at the same time.

Our observations elicit several thoughts and questions. Is it important if patients and physicians rate disease activity differently? After all, patients must look at their disease from their subjective point of view, while physicians presumably judge disease activity from their experience of treating patients with RA. At the same time, is it misleading that patients draw this clear-cut line between RA disease activity and other comorbid conditions, let alone RA joint damage or permanent disability related to RA? Should patients be better informed on which aspects of their disease lead to joint damage and irreversible disability rather than “just” focusing on their feelings at a single point in time? Do physicians need education on how to understand their patients' perceptions better (34), or do patients need more education on how physicians interpret these measures (35, 36)? Perhaps identification of the patient's educational needs will answer the question of how educational programs need to be designed to achieve long-term effects (37). Is there a need to study variations in phrasing of the questions related to the patient's and physician's global assessments in more detail than has hitherto been done? Should focus groups comprising patients and physicians separately and together be formed to foster mutual understanding?

Considering that the shared decision process is so important in chronic disease, these questions are highly pertinent. We conclude from our data that when treating RA, it is important to understand the patient's priorities and problems, many of which will not directly relate to the joint disease, but may arise from underlying comorbidities or chronic damage. This relates especially to the treatment of pain, noninflammatory pain in particular. Once the components of noninflammatory pain are effectively reduced, the patient's and physician's perceptions may become much more similar. The decision about the necessity or lack of necessity of instituting additional disease-modifying treatments can then be made together. Under these conditions, the current ways of assessing disease activity by using composite measures will also more precisely reflect “true” disease activity.

Many of the questions stated above cannot be answered by a study using prospectively acquired, but historically analyzed, routine data. The phrasing of the question for the PGA needs particular emphasis, but there is also a clear responsibility by the evaluator to explain to the patient what a global assessment means. The patient's rating will likely vary considerably according to this explanation. Our study population was rather heterogeneous in terms of cultural background, with many patients being from Turkey, the Balkan States, or Eastern Europe. Pain is expressed differently depending on cultural background (38), which could also influence some of the associations seen in our study. While this heterogeneity expands the scope of the study as compared to patients from a single ethnic background, this aspect may nevertheless affect the generalizability of our findings to other ethnic populations. The strength of our study, however, lies in the fact that it is based on complete clinical, laboratory, and functional data from a large observational database. Many prospective data sets include more patients but lack these complete core set measures (e.g., joint counts, patient's and physician's global assessments, and levels of acute-phase reactants at every visit).

Using current methods (current phrasing, with little explanation given to the patient about the meaning of the different measures), the best way to reduce discrepancies between the PGA and the EGA is probably to assess disease activity with composite measures (39). Ideally, such composite measures should capture the main domains of RA disease activity, that is, the patient's perspective (expressed by pain levels or the PGA), the physician's judgment (such as the findings of joint examinations), as well as laboratory measures (levels of acute-phase reactants) (40). The fact that items from the same domain naturally have a strong intercorrelation underline our findings that pain mainly loads on the PGA and swollen joints on the EGA.

In conclusion, the results of our study call for an awareness that the patient's and the physician's general perceptions of disease activity are drawn from different perspectives. It is important for physicians to understand why patients are drawing the line on the visual analog scale at the place where they actually draw it. In addition, our findings represent a call to treat noninflammatory pain, as this may create better consensus about the course of treatment.


  1. Top of page
  2. Abstract
  7. Acknowledgements

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Aletaha had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Studenic, Radner, Smolen, Aletaha.

Acquisition of data. Studenic, Radner, Smolen, Aletaha.

Analysis and interpretation of data. Studenic, Smolen, Aletaha.


  1. Top of page
  2. Abstract
  7. Acknowledgements

We are indebted to Michael Reiter and Marion Skobek for their vast contributions to the database integrity, and to the biometrics team at our outpatient clinic.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • 1
    Scott DL, Symmons DP, Coulton BL, Popert AJ. Long-term outcome of treating rheumatoid arthritis: results after 20 years. Lancet 1987; 1: 110811.
  • 2
    Scott DL, Pugner K, Kaarela K, Doyle DV, Woolf A, Holmes J, et al. The links between joint damage and disability in rheumatoid arthritis. Rheumatology (Oxford) 2000; 39: 12232.
  • 3
    Welsing PM, van Gestel AM, Swinkels HL, Kiemeney LA, van Riel PL. The relationship between disease activity, joint destruction, and functional capacity over the course of rheumatoid arthritis. Arthritis Rheum 2001; 44: 200917.
  • 4
    Isomaki H. Long-term outcome of rheumatoid arthritis. Scand J Rheumatol Suppl 1992; 95: 38.
  • 5
    Mitchell DM, Spitz PW, Young DY, Bloch DA, McShane DJ, Fries JF. Survival, prognosis, and causes of death in rheumatoid arthritis. Arthritis Rheum 1986; 29: 70614.
  • 6
    Smolen JS, Aletaha D, Bijlsma JW, Breedveld FC, Boumpas D, Burmester G, et al, for the T2T Expert Committee. Treating rheumatoid arthritis to target: recommendations of an international task force. Ann Rheum Dis 2010; 69: 6317.
  • 7
    Ortendahl M. Models based on value and probability in health improve shared decision making. J Eval Clin Pract 2008; 14: 7147.
  • 8
    Berkanovic E, Hurwicz ML, Lachenbruch PA. Concordant and discrepant views of patients' physical functioning. Arthritis Care Res 1995; 8: 94101.
  • 9
    Suarez-Almazor ME, Conner-Spady B, Kendall CJ, Russell AS, Skeith K. Lack of congruence in the ratings of patients' health status by patients and their physicians. Med Decis Making 2001; 21: 11321.
  • 10
    Felson DT, Anderson JJ, Boers M, Bombardier C, Chernoff M, Fried B, et al. The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. Arthritis Rheum 1993; 36: 72940.
  • 11
    Boers M, Tugwell P, Felson DT, van Riel PL, Kirwan JR, Edmonds JP, et al. World Health Organization and International League of Associations for Rheumatology core endpoints for symptom modifying antirheumatic drugs in rheumatoid arthritis clinical trials. J Rheumatol Suppl 1994; 41: 869.
  • 12
    Tugwell P, Boers M, Baker P, Wells G, Snider J. Endpoints in rheumatoid arthritis. J Rheumatol Suppl 1994; 42: 28.
  • 13
    Felson DT, Smolen JS, Wells G, Zhang B, van Tuyl LH, Funovits J, et al. American College of Rheumatology/European League Against Rheumatism provisional definition of remission in rheumatoid arthritis for clinical trials. Arthritis Rheum 2011; 63: 57386.
  • 14
    Vermeer M, Kuper H, van der Biijl A, Baan H, Posthumus M, Brus H, et al. The new ACR/EULAR definition of remission in rheumatoid arthritis: is the patient global assessment criterion too strict? [abstract]. Ann Rheum Dis 2011; 70: 178.
  • 15
    Pincus T, Yazici Y, Bergman M, Swearingen C, Harrington T. A proposed approach to recognise “near-remission” quantitatively without formal joint counts or laboratory tests: a patient self-report questionnaire Routine Assessment of Patient Index Data (RAPID) score as a guide to a “continuous quality improvement” strategy. Clin Exp Rheumatol 2006; 24 Suppl 43: S-60S-675.
  • 16
    Pincus T, Yazici Y, Bergman M, Maclean R, Harrington T. A proposed continuous quality improvement approach to assessment and management of patients with rheumatoid arthritis without formal joint counts, based on quantitative Routine Assessment of Patient Index Data (RAPID) scores on a Multidimensional Health Assessment Questionnaire (MDHAQ). Best Pract Res Clin Rheumatol 2007; 21: 789804.
  • 17
    Kumar K, Gordon C, Barry R, Shaw K, Horne R, Raza K. ‘It’s like taking poison to kill poison but I have to get better': a qualitative study of beliefs about medicines in rheumatoid arthritis and systemic lupus erythematosus patients of South Asian origin. Lupus 2011; 20: 83744.
  • 18
    Aletaha D, Smolen JS. The rheumatoid arthritis patient in the clinic: comparing more than 1,300 consecutive DMARD courses. Rheumatology (Oxford) 2002; 41: 136774.
  • 19
    Aletaha D, Smolen JS. Effectiveness profiles and dose dependent retention of traditional disease modifying antirheumatic drugs for rheumatoid arthritis: an observational study. J Rheumatol 2002; 29: 16318.
  • 20
    Kapral T, Dernoschnig F, Machold KP, Stamm T, Schoels M, Smolen JS, et al. Remission by composite scores in rheumatoid arthritis: are ankles and feet important? Arthritis Res Ther 2007; 9: R72.
  • 21
    Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 31524.
  • 22
    Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham CO III, et al. 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis Rheum 2010; 62: 256981.
  • 23
    Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980; 23: 13745.
  • 24
    Tugwell P, Boers M, on behalf of the OMERACT Committee. Developing consensus on preliminary core efficacy endpoints for rheumatoid arthritis clinical trials. J Rheumatol 1993; 20: 5556.
  • 25
    Aletaha D, Smolen J. The Simplified Disease Activity Index (SDAI) and the Clinical Disease Activity Index (CDAI): a review of their usefulness and validity in rheumatoid arthritis. Clin Exp Rheumatol 2005; 23 Suppl 39: S1008.
  • 26
    Drossaers-Bakker KW, de Buck M, van Zeben D, Zwinderman AH, Breedveld FC, Hazes JM. Long-term course and outcome of functional capacity in rheumatoid arthritis: the effect of disease activity and radiologic damage over time. Arthritis Rheum 1999; 42: 185460.
  • 27
    Dougados M, Ripert M, Hilliquin P, Fardellone P, Brocq O, Brault Y, et al. The influence of the definition of patient global assessment in assessment of disease activity according to the Disease Activity Score (DAS28) in rheumatoid arthritis. J Rheumatol 2011; 38: 23268.
  • 28
    Ward MM, Leigh JP. The relative importance of pain and functional disability to patients with rheumatoid arthritis. J Rheumatol 1993; 20: 14949.
  • 29
    Heiberg T, Kvien TK. Preferences for improved health examined in 1,024 patients with rheumatoid arthritis: pain has highest priority. Arthritis Rheum 2002; 47: 3917.
  • 30
    Aletaha D, Machold KP, Nell VP, Smolen JS. The perception of rheumatoid arthritis core set measures by rheumatologists: results of a survey. Rheumatology (Oxford) 2006; 45: 11339.
  • 31
    Rohekar G, Pope J. Test-retest reliability of patient global assessment and physician global assessment in rheumatoid arthritis. J Rheumatol 2009; 36: 217882.
  • 32
    Nicolau G, Yogui MM, Vallochi TL, Gianini RJ, Laurindo IM, Novaes GS. Sources of discrepancy in patient and physician global assessments of rheumatoid arthritis disease activity. J Rheumatol 2004; 31: 12936.
  • 33
    Khan NA, Spencer HJ, Abda E, Aggarwal A, Alten R, Ancuta C, et al. Determinants of discordance in patients' and physicians' rating of rheumatoid arthritis disease activity. Arthritis Care Res (Hoboken) 2012; 64: 20614.
  • 34
    Van den Bemt BJ, den Broeder AA, van den Hoogen FH, Benraad B, Hekster YA, van Riel P, et al. Making the rheumatologist aware of patients' non-adherence does not improve medication adherence in patients with rheumatoid arthritis. Scand J Rheumatol 2011; 40: 1926.
  • 35
    Combe B, Landewe R, Lukas C, Bolosiu HD, Breedveld F, Dougados M, et al. EULAR recommendations for the management of early arthritis: report of a task force of the European Standing Committee for International Clinical Studies Including Therapeutics (ESCISIT). Ann Rheum Dis 2007; 66: 3445.
  • 36
    Riemsma RP, Kirwan JR, Taal E, Rasker HJ. Patient education for adults with rheumatoid arthritis. Cochrane Database Syst Rev 2003; CD003688.
  • 37
    Meesters JJ, Vliet Vlieland TP, Hill J, Ndosi ME. Measuring educational needs among patients with rheumatoid arthritis using the Dutch version of the Educational Needs Assessment Tool (DENAT) [published erratum appears in Clin Rheumatol 2009;28:1357]. Clin Rheumatol 2009; 28: 10737.
  • 38
    Anderson KO, Green CR, Payne R. Racial and ethnic disparities in pain: causes and consequences of unequal care. J Pain 2009; 10: 1187204.
  • 39
    Goldsmith CH, Smythe HA, Helewa A. Interpretation and power of a pooled index. J Rheumatol 1993; 20: 5758.
  • 40
    Van der Heijde DM, van 't Hof MA, van Riel PL, Theunisse LA, Lubberts EW, van Leeuwen MA, et al. Judging disease activity in clinical practice in rheumatoid arthritis: first step in the development of a disease activity score. Ann Rheum Dis 1990; 49: 91620.