Change of Psychological Distress and PhysicalDisability in Patients With Rheumatoid ArthritisOver the Last Two DecadesCECILE L. OVERMAN,1 MAUD S. JURGENS,2 ERCOLIE R. BOSSEMA,1 JOHANNES W. G. JACOBS,2

JOHANNES W. J. BIJLSMA,2 AND RINIE GEENEN3

Objective. During the past decades, a more cautious approach with respect to prescribing medication and physicalexercise progressed toward evidence-based guidelines regarding the management of rheumatoid arthritis (RA). Currently,physical activity and other means to improve well-being and functioning are encouraged, and the disease is targetedearlier with more intensive and aggressive pharmacologic treatment. The current study examined whether psychologicaldistress and physical disability in patients with RA reduced over the last 2 decades and whether this is explained by areduction of disease activity.Methods. From 1990–2011, consecutive patients with RA (n � 1,151, age range 17–86 years, 68% female, 62% rheuma-toid factor positive) were monitored at diagnosis and after 3–5 years of treatment (followup). Depressed mood, anxiety,and physical disability were predicted in multiple linear regression analyses by year of assessment, disease activity, andpatient demographics.Results. Over the decades, depressed mood (P � 0.01), anxiety (P � 0.001), and physical disability (P � 0.02) reduced atdiagnosis and within-treatment improvement of anxiety (P � 0.04) and physical disability (P < 0.001) increased.Percentages of patients with depressed mood, anxiety, and physical disability at followup changed from 25%, 23%, and53%, respectively, 2 decades ago to 14%, 12%, and 31%, respectively, currently. After taking account of reduction indisease activity, the decrease in physical disability remained significant (P < 0.001).Conclusion. Over the last 2 decades, psychological distress and physical disability decreased. This favorable trend mightpartly be due to reduced disease activity. The results indicate that patients with RA have a better opportunity to live avalued life currently than 20 years ago.

INTRODUCTION

Rheumatoid arthritis (RA) imposes a considerable threat tophysical functioning and psychological well-being (1–4).However, over the past decades, treatment of RA has im-proved considerably, at least in wealthy countries (5), with

possible positive implications for patients’ quality of life.Pharmacologic treatment strategies have shifted from in-tensifying treatment in case of exacerbations to earlier (6)and more intensive treatment (7–10), including combina-tion therapy (11), tight control and treat-to-target strategies(12), and biologic agents (13–16). Furthermore, behavioralrecommendations have shifted from advising rest (17,18)to recommending physical activity (19) and encouragingpatients to live a less restricted life in order to improvewell-being and keep psychological distress at bay (20). Inclinical trials, it has been shown that the new treatmentstrategies are associated with lower levels of disease activ-ity, improved psychological well-being, and better physi-cal functioning (8,9,12,19,21), and population studiesshowed a decrease of joint deformities over the years(22,23).

Over the past 20 years, the enhanced focus on well-beingand functioning and the improved pharmacotherapy mayhave resulted in a reduction of psychological distress andphysical disability. This has only been indicated for ashorter time period in studies with few cross-sectional

Supported by a grant from the Faculty of Social andBehavioral Sciences of Utrecht University and by unre-stricted grants from the Dutch Arthritis Foundation (LLP-17).

1Cecile L. Overman, MSc, Ercolie R. Bossema, PhD:Utrecht University, Utrecht, The Netherlands; 2Maud S.Jurgens, MSc, Johannes W. G. Jacobs, MD, PhD, JohannesW. J. Bijlsma, MD, PhD: University Medical Center Utrecht,Utrecht, The Netherlands; 3Rinie Geenen, PhD: Utrecht Uni-versity and University Medical Center Utrecht, Utrecht, TheNetherlands.

Address correspondence to Cecile L. Overman, MSc, De-partment of Clinical and Health Psychology, Utrecht Uni-versity, (Heidelberglaan 1) PO Box 80.140, 3508 TC Utrecht,The Netherlands. E-mail: C.L.Overman@uu.nl.

Submitted for publication July 16, 2013; accepted in re-vised form October 15, 2013.

Arthritis Care & ResearchVol. 66, No. 5, May 2014, pp 671–678DOI 10.1002/acr.22211© 2014, American College of Rheumatology

ORIGINAL ARTICLE

671

cohorts, not extending to recent years (24–28). The aim ofthe current study was to examine whether psychologicaldistress and physical disability decreased over the last 2decades and whether this decrease was associated with aparallel reduction of disease activity.

MATERIALS AND METHODS

Participants. From 1990–2011, a total of 1,151 patientswith recent RA (median 47 patients per year, range 10–104per year) were repeatedly assessed from the time of diag-nosis until 5 years thereafter. Patients were recruited inrheumatology departments in The Netherlands, collaborat-ing in the Utrecht Rheumatoid Arthritis Cohort studygroup. They gave written informed consent to participatein one of several prospective trials comparing the effec-tiveness of different drug treatment strategies, includingstrategies that were conventional at the time. These studieswere approved by the ethical review boards of the partic-ipating hospitals. Details have been described elsewhere(7–9,12). Briefly, inclusion criteria were a diagnosis of RAaccording to the American College of Rheumatology crite-ria for RA (29), disease duration �1 year, and age �16years. Excluded from the study were patients diagnosedwith psychiatric disorders, patients with comorbid dis-eases, patients with problematic drug usage, and fertilefemale patients not taking adequate contraceptive mea-sures or who were pregnant or breastfeeding.

Assessments. Included in this study were assessmentsat diagnosis and after 4 years. If no data were available atthe fourth followup year, data gathered at the fifth or, ifalso not available, the third followup year were used. Thistimeframe guaranteed a followup period with conven-tional treatment of at least 1 year, since most patientsparticipated in 2-year medical trials. Data at diagnosiswere collected from 1990–2008. Followup data were ac-quired from 1993–2011.

Psychological distress was assessed with the depressedmood and anxiety scales of the Impact of Rheumatic Dis-eases on General Health and Lifestyle (IRGL) questionnaire

(30,31). The depressed mood scale (range 0–24) consists of6 items derived from a questionnaire by Zwart and Spoor-en (32). The anxiety scale (range 10–40) consists of 10items derived from the Spielberger State-Trait AnxietyInventory (33,34). Functional status was assessed with thedisability index of the Health Assessment Questionnaire(HAQ) (35,36). The disability index (range 0–3) comprises20 items representing difficulty in performing activities in8 areas of daily living: dressing and grooming, arising,eating, walking, hygiene, reach, grip, and common dailyactivities. Both the items of the IRGL and the HAQ arescored on a 4-point Likert scale, with higher scores repre-senting more depressed mood, a higher level of anxiety,and more disability in daily functioning, respectively. Cut-off scores representing clinically relevant levels were �6for depressed mood and �23 for anxiety (31). A cutoff of�1 was used to reflect a clinical level of physical disability(37,38). Reliability and validity of the IRGL and the HAQare satisfactory (30,36,39).

Disease activity was assessed with the erythrocyte sed-imentation rate (ESR) in mm/hour (Westergren) and theThompson Articular Index (range 0–534), a weightedscore including both swollen and painful joints (40). TheThompson joint score was adjusted because for 448 pa-tients, only the sum of counted swollen and painful fingerand toe joints had been recorded at diagnosis. To calculatea Thompson joint score for these patients, it was assumedthat when, for example, 3 finger joints were swollen and 2finger joints were painful, 2 finger joints were both swollenand painful. Furthermore, involvement of the big toe (nor-mally assigned 8 points) could not be inferred for thesepatients. Therefore, for all patients, 3 points were assignedto any swollen and painful toe, resulting in an adjustedtotal score range of 0–524.

Statistical analysis. Established cutoff scores (31,37,38)were used to describe the number of patients with a clin-ical level of depressed mood, anxiety, and physical dis-ability across the 2 decades.

To examine whether psychological distress and physicaldisability improved over the past 2 decades, multiple lin-ear regression analyses were applied. The variables wereentered hierarchically in blocks, using a fixed order. De-pendent variables were IRGL depressed mood, IRGL anx-iety, and the HAQ disability index at diagnosis and base-line-adjusted scores at followup. The regression modelswere built up in 3 blocks (models), as described below.

Year of assessment was entered as a predictor in model1 to examine time trends over the 2 decades. Polynomialswere tested, but did not improve the fit of a linear modeland were not included in final regression analyses.

ESR and joint score were added in model 2 as possiblemediators of the time trends of psychological distress andphysical disability over the 2 decades. When analysesshowed that both year of assessment and ESR or joint scoresignificantly predicted psychological distress or physicaldisability, and also that year of assessment significantlypredicted ESR or joint score, mediation of the time trendsby disease activity variables was examined by calculating

Significance & Innovations● Whereas 20 years ago still 2 of 4 patients with

newly diagnosed arthritis were physically dis-abled after the first 4 years of treatment, currentlythis holds for 1 of 4 patients; in addition, thepercentage of patients with depressed mood andanxiety reduced by �50%.

● The favorable trend of physical disability, de-pressed mood, and anxiety decreasing over the last2 decades may partly be due to reduced diseaseactivity.

● Our research findings suggest that currently it iseasier to live a valued life while having rheuma-toid arthritis than 20 years ago.

672 Overman et al

bootstrap 95% confidence intervals (95% CIs) in multiplemediator models based on 1,000 samples (41).

Age, sex, education level, marital status, and rheuma-toid factor (RF) status were included in model 3 to exam-ine whether the time trends of psychological distress andphysical disability remained after taking these variables

into account. This third step was added to the regressionmodels when it became clear that the composition of pa-tient influx at diagnosis and followup changed over theyears. This was examined by predicting age, sex, educa-tion level, marital status, and RF status by year of assess-ment in linear and logistic regression analyses.

For all linear regression analyses, standardized coeffi-cients (�) are reported to compare the effects of the indi-vidual predictors within a model and the (added) varianceexplained by each of the models (�R2) to see if addingvariables improved the fit of the overall model. Cohen’s dwas used to express the magnitude of change in the de-pendent variables over the 2 decades, with values of 0.2,0.5, and 0.8 representing cutoffs of small, moderate, andlarge differences, respectively. For the calculation of Co-hen’s d, model estimates of mean values and overall SDswere used. For the logistic regression analyses, odds ratios(ORs) and goodness of fit of the models (Nagelkerke’s R2)are reported. In all analyses, significance levels were set atP less than 0.05. Data were analyzed using SPSS, version20.

RESULTS

Characteristics of the study population. The character-istics of the patients are shown in Table 1. Of the 1,151patients, 1,060 patients provided the necessary data atdiagnosis and 768 patients provided data at both diagnosisand followup. Patients with (n � 292) and without missingdata at followup did not differ on characteristics (P � 0.07for all). Patients with missing data at diagnosis (n � 91)were more often male (P � 0.003) and more often had amiddle (P � 0.02) or high level of education (P � 0.001);

Figure 1. Percentage of patients with depressed mood, anxiety, or physical disability at diagnosis and followup. Percentages shown arethose of patients with a diagnosis between 1990 and 2007 and followup scores between 1994 and 2011. An Impact of Rheumatic Diseaseson General Health and Lifestyle depressed mood score �6 and anxiety score �23 indicated psychological distress and a Health AssessmentQuestionnaire disability index score �1 indicated physical disability.

Table 1. Characteristics of the 1,151 patients withrheumatoid arthritis*

Value

Female sex 780 (68)Age, mean � SD years 55 � 15

Males 57 � 13Females 54 � 15

Education levelLow 143 (12)Middle 560 (49)High 246 (21)Unknown 202 (18)

Marital statusSingle 111 (10)Married or cohabiting 827 (72)Divorced 57 (5)Widowed 116 (10)Unknown 40 (3)

Rheumatoid factor statusPositive 715 (62)Negative 372 (32)Unknown 64 (6)

* Values are the number (percentage) unless indicated otherwise.Low education level � primary school or lower vocational second-ary education; middle education level � intermediate general sec-ondary education (high school) or intermediate vocational educa-tion; high education level � higher general secondary education(high school), higher vocational education, or university education.

Two Decades of Psychological and Physical Health Improvement in RA 673

they did not differ on other characteristics (P � 0.27 forall).

Figure 1 displays the course of improvement across the2 decades by showing the yearly percentages of patientswith psychological distress or physical disability at diag-nosis and followup. Psychological distress and physicaldisability were common in our patient sample and moreprevalent at diagnosis than at followup. At diagnosis, theaverage percentage of patients with depressed mood, anx-iety, or physical disability declined from 43%, 34%, and64%, respectively, in 1990–1994 to 32%, 21%, and 60%,respectively, in 2004–2008. At followup, these percent-ages were 25%, 23%, and 53%, respectively, in 1994–1998 and 14%, 12%, and 31%, respectively, in 2007–2011.

Time trends of psychological distress and physical dis-ability (model 1). The results of multiple linear regressionanalyses at diagnosis are shown in Table 2. Model 1 showsthat, across the 2 decades, levels of depressed mood (P �0.01), anxiety (P � 0.001), and physical disability (P �0.02) at diagnosis decreased. The decrease was small fordepressed mood (d � �0.4) and physical disability (d ��0.3) and moderate for anxiety (d � �0.5).

The baseline-adjusted prediction of psychological dis-tress and physical disability at followup is shown in Table3. Having entered values at diagnosis in the first step, nowwithin-treatment change (instead of absolute followup lev-

els) in psychological distress and physical disabilityacross the 2 decades was predicted. Model 1 shows thatthe within-treatment improvement of anxiety (P � 0.04)and physical disability (P � 0.001), but not of depressedmood (P � 0.32), increased across the 2 decades. Thisincrease in within-treatment improvement across the de-cades was small for anxiety (d � �0.2) and moderate tolarge for physical disability (d � �0.7).

Mediation of time trends by disease activity (model 2).At diagnosis, after having taken account of time trends,disease activity was significantly associated with de-pressed mood (�R2 � 0.02, P � 0.001), anxiety (�R2 �0.02, P � 0.001), and physical disability (�R2 � 0.23, P �0.001) (Table 2). To test mediation, we further observedthat ESR (P � 0.02) and joint score (P � 0.007) at diagnosisdecreased over the 2 decades; both decreases were small(d � �0.3 and d � �0.4, respectively). Bootstrap 95% CIsshowed that at diagnosis, the decrease in ESR was a partialmediator of the decrease in depressed mood (95% CI�0.023, �0.001) and physical disability (95% CI �0.007,�0.001), and that the decrease in joint score was a partialmediator of the decrease in anxiety (95% CI �0.032,�0.002) and physical disability (95% CI �0.007, �0.001).

At followup, after having taken account of the values atdiagnosis of psychological distress and physical disabilityand time trends, disease activity was significantly associ-

Table 2. Multiple linear regression analyses at diagnosis*

Depressedmood Anxiety

Physicaldisability

� �R2 � �R2 � �R2

Model 1 0.01† 0.01‡ 0.01†Year �0.09† �0.12‡ �0.08†

Model 2 0.02‡ 0.02§ 0.23§Year �0.08† �0.10‡ �0.03ESR 0.11‡ 0.07 0.30§Joint score 0.06 0.10‡ 0.28§

Model 3 0.03§ 0.03§ 0.04§Year �0.04 �0.07 0.01ESR 0.12‡ 0.08† 0.29§Joint score 0.05 0.10† 0.27§Age �0.11‡ �0.11† 0.10‡Sex¶ 0.09† 0.07 0.10‡Education level#

Middle �0.06 �0.07 �0.03High �0.15‡ �0.10 �0.08

Marital status**Married/cohabiting 0.07 0.01 0.12†Divorced/widowed 0.12† 0.14† 0.11†

RF status†† 0.02 �0.01 �0.06†

* � � standardized regression coefficient; �R2 � the additional change in the proportion of varianceexplained by the model compared to the previous model (the very first �R2 value is the change inexplained variance between model 1 and a model without any predictors); ESR � erythrocyte sedimen-tation rate; joint score � Thompson joint score (adjusted range 0–524); RF � rheumatoid factor.† P � 0.05.‡ P � 0.01.§ P � 0.001.¶ Male � 0, female � 1.# Represented by 2 dummy variables with the lowest level of education as the reference category.** Represented by 2 dummy variables with single status as the reference category.†† RF-negative status � 0, RF-positive status � 1.

674 Overman et al

ated with baseline-adjusted depressed mood (�R2 � 0.03,P � 0.001), anxiety (�R2 � 0.02, P � 0.001), and physicaldisability (�R2 � 0.10, P � 0.001) (Table 3). To test medi-ation, we first observed that ESR (P � 0.008) and jointscore (P � 0.001) at followup decreased over the 2 de-cades. These decreases were small (d � �0.3) and moder-ate to large (d � �0.7), respectively. Bootstrap 95% CIsshowed that at followup, the decrease in ESR was a partialmediator of the baseline-adjusted decrease in physical dis-ability (95% CI �0.004, �0.001) and the decrease in jointscore was a partial mediator of the baseline-adjusted de-crease in anxiety (95% CI �0.065, �0.010) and physicaldisability (95% CI �0.011, �0.004).

Changes in the composition of patient influx (model 3).In the group of patients who provided data at diagnosis(n � 1,060), age decreased over the 2 decades (d � �0.40,R2 � 0.01, � � �0.10, P � 0.001), the number of patientswith a middle or high level of education increased(Nagelkerke’s R2 � 0.15, ORs 1.24 and 1.38, respectively,P � 0.001), the number of patients with a positive RFstatus increased (Nagelkerke’s R2 � 0.01, OR 1.03, P �

0.04), and the number of divorced or widowed patientsdecreased (Nagelkerke’s R2 � 0.01, OR 0.92, P � 0.005).Other patient characteristics did not change over the years(P � 0.26 for all) for this group. In the group of patientswho provided data at both diagnosis and followup (n �768), the number of patients with a middle or high level ofeducation increased (Nagelkerke’s R2 � 0.15, ORs 1.23 and1.37, respectively, P � 0.001). No other changes in patientcharacteristics were observed across the years for thisgroup (P � 0.11 for all).

Model 3 tested whether the time trends of psychologicaldistress and physical disability were affected by thechange in the composition of patient influx. Patient char-acteristics were inconsistently associated with at least oneof the psychological distress or physical disability vari-ables at diagnosis or followup. Adding all patient charac-teristics to the model eliminated the significance of year ofassessment as a predictor of depressed mood and anxietyat diagnosis (Table 2), whereas the baseline-adjusted fol-lowup decrease in physical disability as a function of yearof assessment remained significant (Table 3).

Table 3. Baseline-adjusted multiple linear regression analyses at followup*

Depressedmood Anxiety

Physicaldisability

� �R2 � �R2 � �R2

Model 1 0.001 0.01† 0.04‡Baseline value 0.44‡ 0.52‡ 0.46‡Year �0.04 �0.07† �0.21‡

Model 2 0.03‡ 0.02‡ 0.10‡Baseline value 0.44‡ 0.51‡ 0.44‡Year 0.001 �0.04 �0.14‡ESR 0.06 0.05 0.15‡Joint score 0.15‡ 0.13‡ 0.24‡

Model 3 0.02† 0.03‡ 0.04‡Baseline value 0.42‡ 0.50‡ 0.41‡Year 0.05 0.01 �0.11§ESR 0.05 0.02 0.14‡Joint score 0.15‡ 0.12‡ 0.24‡Age 0.04 0.08† 0.04Sex¶ 0.07 0.08† 0.14‡Education level#

Middle �0.09 �0.11† 0.05High �0.15† �0.16§ �0.07

Marital status**Married/cohabiting 0.02 �0.01 0.03Divorced/widowed 0.03 �0.02 0.05

RF status†† �0.003 0.09§ 0.03

* � � standardized regression coefficient; �R2 � change in the additional proportion of variance explained by the modelcompared to the previous model (the very first �R2 value is the change in explained variance between model 1 and a model withonly the baseline value as a predictor); baseline value � baseline value of the dependent variable; ESR � erythrocytesedimentation rate; joint score � Thompson joint score (adjusted range 0–524); RF � rheumatoid factor.† P � 0.05.‡ P � 0.001.§ P � 0.01.¶ Male � 0, female � 1.# Represented by 2 dummy variables with the lowest level of education as the reference category.** Represented by 2 dummy variables with single status as the reference category.†† RF-negative status � 0, RF-positive status � 1.

Two Decades of Psychological and Physical Health Improvement in RA 675

DISCUSSION

The current study in patients with a recent RA diagnosisshowed that psychological distress and physical disabilitydecreased across the last 2 decades. Reduction in diseaseactivity partly explained this decrease. Across the de-cades, the trend of progressively decreasing physical dis-ability after the first years of treatment especially wasnoteworthy, and this favorable trend remained significantafter having taken account of the reduction of diseaseactivity and change in the composition of patient influx.

Adding to findings from previous studies (24,28), ourstudy examined whether parallel changes in outcomeswere associated. Our study suggests that the favorabletrends of psychological distress and physical disabilitymight be caused in part by reduced disease activity. Bothimproved pharmacologic treatment, as suggested previ-ously (23), and improved nonpharmacologic managementcould have induced the favorable trends of psychologicaldistress and physical disability observed after the firstyears of treatment by reducing disease activity across thelast 2 decades.

Improved treatment strategies could have affected psy-chological distress and physical disability by reducingdisease activity, but also by other mechanisms. Targetednonpharmacologic therapies in small samples have shownpositive effects of exercise therapy (19) and cognitive–behavioral therapy (42), but much more important in thelarge population of patients with a recent RA diagnosismight have been the changed education by rheumatolo-gists and health professionals enhancing physical activityand encouraging patients to live a valued life (20). Further-more, since proinflammatory cytokines may have behav-ioral effects (43,44), the blocking of these cytokines withbiologic agents in recent years may have caused a reduc-tion of psychological distress and physical disability notfully explained by reduced disease activity.

The improvements of psychological distress and physi-cal disability and reduced disease activity at diagnosiscannot be ascribed to improved treatment strategies. Theseimprovements might be due to earlier diagnosis over theyears (6), as suggested in our study by a decrease in age atthe time of diagnosis across the past 20 years.

Patient characteristics changed across the years in ourpatient sample. The observed increase in education levelduring the past 20 years is not specific for our patientsample, but has also been found in the general population(45,46), accompanied by an increase in socioeconomicstatus (46) and life expectancy (47). When adjusting forhistorical trends in analysis, as we did for education level,it is impossible to ascertain whether these trends play acausal role in the observed psychological and functionalimprovements across the decades, or whether they onlyreflect a historical change occurring simultaneously, af-fecting the improvements statistically but not in real life.In any case, our observed trend of progressively decreasingphysical disability after the first years of treatment re-mained significant after correction for historical changesin patient characteristics.

Improved treatment strategies have shown to be capableof improving patients’ psychological well-being and phys-

ical functioning in randomized controlled trials (8,9,12,19,21). Therefore, treatment improving over the decades isa likely candidate to explain, at least in part, the improve-ment in psychological well-being and physical function-ing. However, our statistical tests do not give a final an-swer. It cannot be excluded that trends of improvingtreatment, decreasing disease activity, and increasingwell-being and functioning are merely parallel occur-rences without the existence of a causal connection.

The decrease in psychological distress over the decadeswas not as pronounced as the decrease in physical disabil-ity. However, our finding that mood and anxiety improvedacross the decades, even with small effect sizes, is remark-able when taking into consideration that in the past de-cades, psychological distress increased instead of de-creased in the general population (48).

The current study has strengths that add to previousfindings (23–28,37,49). It is one of few population studies(25–28) examining time trends of both physical function-ing and psychological well-being, and the first to examinereduced disease activity as a possible mediator of thesetime trends. Moreover, time trends were examined withannual cohorts over a long period of time (21 years), add-ing to findings of population studies based on a morelimited number of cross-sectional cohorts and coveringfewer years (25–28). Also, our patient sample had a uni-form disease duration. Therefore, in contrast with othercross-sectional population studies (24–27), the effect ofyear of assessment was not confounded by disease dura-tion (50). Lastly, our data covered more recent years thanprevious population studies (23–28,37,49); therefore, ef-fects of the newest treatments were included.

The current study also has limitations. The results donot refer to specific treatment modalities; all patients re-ceived treatment according to guidelines at that time. Fur-thermore, our findings do not generalize beyond patientswith a recent RA diagnosis or beyond wealthy countries,since the clinical burden of RA is still especially substan-tial in less prosperous countries (5). Lastly, it was notpossible to definitively establish reduced disease activityas a cause of reduced psychological distress and physicaldisability or to examine prediction or confounding of theresults by patient characteristics that changed over thedecades, such as increased level of education, which is alimitation inherent to the type of study design.

In conclusion, in this population of patients with arecent RA diagnosis, psychological distress and physicaldisability reduced significantly over the last 2 decades.This favorable trend might partly be due to a decrease indisease activity. The research findings indicate that it iseasier to live a valued life while having RA currently than20 years ago.

ACKNOWLEDGMENTSThe authors would like to thank all of the participatingrheumatologists and research nurses of the Utrecht Rheu-matoid Arthritis Cohort study group for their help in datacollection.

676 Overman et al

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising itcritically for important intellectual content, and all authors ap-proved the final version to be published. Ms Overman had fullaccess to all of the data in the study and takes responsibility forthe integrity of the data and the accuracy of the data analysis.Study conception and design. Overman, Jurgens, Geenen.Acquisition of data. Bijlsma.Analysis and interpretation of data. Overman, Jurgens, Bossema,Jacobs, Geenen.

REFERENCES

1. Dickens C, McGowan L, Clark-Carter D, Creed F. Depressionin rheumatoid arthritis: a systematic review of the literaturewith meta-analysis. Psychosom Med 2002;64:52–60.

2. VanDyke MM, Parker JC, Smarr KL, Hewett JE, Johnson GE,Slaughter JR, et al. Anxiety in rheumatoid arthritis. ArthritisRheum 2004;51:408–12.

3. Isik A, Koca SS, Ozturk A, Mermi O. Anxiety and depressionin patients with rheumatoid arthritis. Clin Rheumatol 2007;26:872–78.

4. Overman CL, Bossema ER, van Middendorp H, Wijngaards-deMeij L, Verstappen SM, Bulder M, et al. The prospectiveassociation between psychological distress and disease activ-ity in rheumatoid arthritis: a multilevel regression analysis.Ann Rheum Dis 2012;71:192–7.

5. Sokka T, Kautiainen H, Pincus T, Toloza S, da Rocha CastelarPinheiro G, Lazovskis J, et al. Disparities in rheumatoid ar-thritis disease activity according to gross domestic product in25 countries in the QUEST-RA database. Ann Rheum Dis2009;68:1666–72.

6. Emery P, Breedveld FC, Dougados M, Kalden JR, Schiff MH,Smolen JS. Early referral recommendation for newly diag-nosed rheumatoid arthritis: evidence based development of aclinical guide. Ann Rheum Dis 2002;61:290–7.

7. Van der Heide A, Jacobs JW, Bijlsma JW, Heurkens AH, vanBooma-Frankfort C, van der Veen MJ, et al. The effectivenessof early treatment with “second-line” antirheumatic drugs: arandomized, controlled trial. Ann Intern Med 1996;124:699–707.

8. Verstappen SM, Jacobs JW, Bijlsma JW, Heurkens AH, vanBooma-Frankfort C, ter Borg EJ, et al, on behalf of the UtrechtArthritis Cohort Study Group. Five-year follow-up of rheuma-toid arthritis patients after early treatment with disease-mod-ifying antirheumatic drugs versus treatment according to thepyramid approach in the first year. Arthritis Rheum 2003;48:1797–807.

9. Verstappen SM, Jacobs JW, van der Veen MJ, Heurkens AH,Schenk Y, ter Borg EJ, et al. Intensive treatment with metho-trexate in early rheumatoid arthritis: aiming for remission.Computer Assisted Management in Early Rheumatoid Arthri-tis (CAMERA, an open-label strategy trial). Ann Rheum Dis2007;66:1443–9.

10. Jurgens MS, Welsing PM, Jacobs JW. Overview and analysis oftreat-to-target trials in rheumatoid arthritis reporting on re-mission. Clin Exp Rheumatol 2012;30:S56–63.

11. Boers M, Verhoeven AC, Markusse HM, van de Laar MA,Westhovens R, van Denderen JC, et al. Randomised compar-ison of combined step-down prednisolone, methotrexate andsulphasalazine with sulphasalazine alone in early rheuma-toid arthritis. Lancet 1997;350:309–18.

12. Bakker MF, Jacobs JW, Welsing PM, Verstappen SM, TekstraJ, Ton E, et al. Low-dose prednisone inclusion in a metho-trexate-based, tight control strategy for early rheumatoidarthritis: a randomized trial. Ann Intern Med 2012;156:329–39.

13. Furst D, Breedveld F, Kalden J, Smolen JS, Burmester GR,Bijlsma JW, et al. Updated consensus statement on biologicalagents, specifically tumour necrosis factor � (TNF�) blockingagents and interleukin-1 receptor antagonist (IL-1ra), for the

treatment of rheumatic diseases, 2005. Ann Rheum Dis 2005;64:iv2–14.

14. Smolen JS, Boers M, Abadie EC, Breedveld FC, Emery P,Bardin T, et al. Recommendations for an update of 2003European regulatory requirements for registration of drugs tobe used in the treatment of RA. Curr Med Res Opin 2011;27:315–25.

15. Jung YO, Kim HA. Recent paradigm shifts in the diagnosisand treatment of rheumatoid arthritis. Korean J Intern Med2012;27:378–87.

16. Sharma P, Pathak K. Are biological targets the final goal forrheumatoid arthritis therapy? Expert Opin Biol Ther 2012;12:1611–22.

17. Smith RD, Polley HF. Rest therapy for rheumatoid arthritis.Mayo Clin Proc 1978;53:141–5.

18. De Ridder D, Geenen R, Kuijer R, van Middendorp H. Psy-chological adjustment to chronic disease. Lancet 2008;372:246–55.

19. De Jong Z, Munneke M, Zwinderman AH, Kroon HM, JansenA, Ronday KH, et al. Is a long-term high-intensity exerciseprogram effective and safe in patients with rheumatoid arthri-tis? Results of a randomized controlled trial. Arthritis Rheum2003;48:2415–24.

20. Geenen R, Mulligan K, Shipley M, Newman S. Psychosocialdimensions of the rheumatic diseases. In: Bijlsma JW, Bur-meister GR, da Silva JA, Faarvang KL, Hachulla E, Mariette X,editors. EULAR compendium on rheumatic diseases. London:BMJ Publishing Group; 2009. p. 635–50.

21. Singh JA, Christensen R, Wells GA, Suarez-Almazor ME,Buchbinder R, Lopez-Olivo MA, et al. A network meta-analy-sis of randomized controlled trials of biologics for rheumatoidarthritis: a Cochrane overview. CMAJ 2009;181:787–96.

22. Sokka T, Kautiainen H, Hakkinen A, Hannonen P. Radio-graphic progression is getting milder in patients with earlyrheumatoid arthritis: results of 3 cohorts over 5 years. J Rheu-matol 2004;31:1073–82.

23. Finckh A, Choi HK, Wolfe F. Progression of radiographic jointdamage in different eras: trends towards milder disease inrheumatoid arthritis are attributable to improved treatment.Ann Rheum Dis 2006;65:1192–7.

24. Pincus T, Sokka T, Kautiainen H. Patients seen for standardrheumatoid arthritis care have significantly better articular,radiographic, laboratory, and functional status in 2000 than in1985. Arthritis Rheum 2005;52:1009–19.

25. Heiberg T, Finset A, Uhlig T, Kvien TK. Seven year changes inhealth status and priorities for improvement of health inpatients with rheumatoid arthritis. Ann Rheum Dis 2005;64:191–5.

26. Soderlin MK, Lindroth Y, Jacobsson LT. Trends in medicationand health-related quality of life in a population-based rheu-matoid arthritis register in Malmö, Sweden. Rheumatology(Oxford) 2007;46:1355–8.

27. Uhlig T, Heiberg T, Mowinckel P, Kvien TK. Rheumatoidarthritis is milder in the new millennium: health status in RApatients 1994-2004. Ann Rheum Dis 2008;67:1710–5.

28. Soderlin MK, Lindroth Y, Turesson C, Jacobsson LT. A moreactive treatment has profound effects on the health status ofrheumatoid arthritis (RA) patients: results from a population-based RA register in Malmö, Sweden, 1997-2005. ScandJ Rheumatol 2010;39:206–11.

29. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,Cooper NS, et al. The American Rheumatism Association1987 revised criteria for the classification of rheumatoid ar-thritis. Arthritis Rheum 1988;31:315–24.

30. Huiskes CJ, Kraaimaat FW, Bijlsma JW. Measuring healthstatus in rheumatoid arthritis: development and refinement ofa Dutch health status instrument. Gedrag Gezond 1990;18:78–89. In Dutch.

31. Huiskes CJ, Kraaimaat FW, Bijlsma JW. IRGL: Impact of Ar-thritis on Health and Lifestyle self-report questionnaire.Nijmegen: University Medical Center St. Radboud Nijmegen;2004. In Dutch.

32. Zwart F, Spooren J. Development of a self-report question-

Two Decades of Psychological and Physical Health Improvement in RA 677

naire for depression. Utrecht: University Medical CenterUtrecht; 1982. In Dutch.

33. Spielberger CD, Gorsuch RL, Lushene RE. Manual of theState-Trait Anxiety Inventory. Palo Alto (CA): Consulting Psy-chologists Press; 1970.

34. Van der Ploeg HM, Defares PB, Spielberger CD. Handleidingbij de ZelfBeoordelingsVragenlijst ZBV. Lisse: Swets & Zeit-linger; 1980.

35. Fries JF, Spitz PW, Young DY. The dimensions of healthoutcomes: the Health Assessment Questionnaire, disabilityand pain scales. J Rheumatol 1982;9:789–93.

36. Bijlsma JW, Oude Heuvel CH, Zaalberg A. Development andvalidation of the Dutch questionnaire capacities of daily life(VDF) for patients with rheumatoid arthritis. J Rehabil Sci1990;3:71–4.

37. Krishnan E, Fries JF. Reduction in long-term functional dis-ability in rheumatoid arthritis from 1977 to 1998: a longitu-dinal study of 3035 patients. Am J Med 2003;115:371–6.

38. Krishnan E, Sokka T, Hakkinen A, Hubert H, Hannonen P.Normative values for the Health Assessment Questionnairedisability index: benchmarking disability in the general pop-ulation. Arthritis Rheum 2004;50:953–60.

39. Van Leeuwen MA, van der Heijde DM, van Rijswijk MH,Houtman PM, van Riel PL, van de Putte LB, et al. Interrela-tionship of outcome measures and process variables in earlyrheumatoid arthritis: a comparison of radiologic damage,physical disability, joint counts, and acute phase reactants.J Rheumatol 1994;21:425–9.

40. Thompson PW, Silman AJ, Kirwan JR, Currey HL. Articularindices of joint inflammation in rheumatoid arthritis: corre-

lation with the acute-phase response. Arthritis Rheum 1987;30:618–23.

41. Preacher KJ, Hayes AF. Asymptotic and resampling strategiesfor assessing and comparing indirect effects in multiple me-diator models. Behav Res Methods 2008;40:879–91.

42. Evers AW, Zautra A, Thieme K. Stress and resilience in rheu-matic diseases: a review and glimpse into the future. Nat RevRheumatol 2011;7:409–15.

43. Dantzer R. Cytokine-induced sickness behavior: mechanismsand implications. Ann N Y Acad Sci 2001;933:222–34.

44. Capuron L, Gumnick JF, Musselman DL, Lawson DH, Reem-snyder A, Nemeroff CB, et al. Neurobehavioral effects of in-terferon-� in cancer patients: phenomenology and paroxetineresponsiveness of symptom dimensions. Neuropsychophar-macology 2002;26:643–52.

45. Beets G, van Nimwegen N. Population issues in The Nether-lands. Rev Popul Soc Policy 2000;9:87–117.

46. OECD iLibrary. Education at a glance 2012: highlights. 2012.URL: http://dx.doi.org/10.1787/eag_highlights-2012-en.

47. Veenhoven R. Is life getting better? Eur Psychol 2005;10:330–43.

48. Twenge JM. The age of anxiety? The birth cohort change inanxiety and neuroticism, 1952-1993. J Pers Soc Psychol 2000;79:1007–21.

49. Welsing PM, Fransen J, van Riel PL. Is the disease course ofrheumatoid arthritis becoming milder? Time trends since1985 in an inception cohort of early rheumatoid arthritis.Arthritis Rheum 2005;52:2616–24.

50. Boonen A, Landewe R. Health status in rheumatoid arthritisover 7 years. Ann Rheum Dis 2005;64:173–5.

678 Overman et al