Measures of rheumatoid arthritis disease activity: Patient (PtGA) and Provider (PrGA) Global Assessment of Disease Activity, Disease Activity Score (DAS) and Disease Activity Score With 28-Joint Counts (DAS28), Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Patient Activity Score (PAS) and Patient Activity Score-II (PASII), Routine Assessment of Patient Index Data (RAPID), Rheumatoid Arthritis Disease Activity Index (RADAI) and Rheumatoid Arthritis Disease Activity Index-5 (RADAI-5), Chronic Arthritis Systemic Index (CASI), Patient-Based Disease Activity Score With ESR (PDAS1) and Patient-Based Disease Activity Score Without ESR (PDAS2), and Mean Overall Index for Rheumatoid Arthritis (MOI-RA)

Authors


INTRODUCTION

Numerous rheumatoid arthritis (RA) disease activity measurement tools are currently available for use. The psychometric data related to these tools have been published over the course of decades and across numerous journals. Consequently, the majority of this information remains inaccessible to practicing and academic rheumatologists alike. To facilitate the availability of this information, the American College of Rheumatology (ACR) RA Clinical Disease Activity Measures Working Group performed a systematic review and compiled the resulting data into an evidence report, which constitutes the majority of this publication. The following psychometric analysis is based on currently published literature with abstract publications excluded. More recent measurement tools with inadequately published data, such as the Global Arthritis Score, may ultimately prove to be equal or superior to those discussed in this article. Time to complete each measure was divided into 1) time for patient completion without assistance from providers; 2) time for provider completion, including the time required to score the measure; and 3) time required for laboratory studies. Our assessment of the Disease Activity Score, Disease Activity Score with 28-joint counts, and Rheumatoid Arthritis Disease Activity Index updates that of Fransen et al, who performed this exercise in 2003 (1).

All RA disease activity measurement tools discussed herein produce a single continuous index of disease activity and were chosen based on expert panel recommendations for use in the clinical setting as well as their current use in the research and clinical settings. Continuous composite indices producing a single score have an advantage over the interpretation of individual components of disease activity as they provide clinically meaningful and reliable estimates of disease activity with interpretation of multiple data points simultaneously and are more responsive to change than single items (2). Continuous indices are less susceptible to selection bias related to the reporting of a single measurement (3) and are also preferable for statistical analysis in studies, with the most commonly used composite indices able to quantify disease activity, even in the lower ends of the scales (4). Additionally, composite indices are recommended by many insurers and regulators to justify escalation of RA therapy (5, 6). On the other hand, simple visual analog scales are widely used and may be the most feasible method for disease activity monitoring in clinical settings. Consequently, the patient global assessment of disease activity and provider global assessment of disease activity will be included in this report, even though these measures may factor in irreversible damage.

A number of measures are not included in this article for various reasons, including lack of use and clinic feasibility; the full list of starting measures and the methods for exclusion are covered in our parallel study. Despite inclusion of a continuous score, the ACR hybrid also requires scoring a change from a prior assessment and will not be discussed, as it is unlikely to be useful at point of care in a clinical setting when a patient may be seen at only a single visit. New criteria for RA remission, endorsed jointly by the ACR and European League Against Rheumatism, were unveiled at the 2010 ACR Annual Scientific Meeting, with formal publication pending. Unfortunately, no publications yet exist comparing the remission cutoffs for each measurement tool to these new criteria. It is important to recognize that for all of the disease activity measurement tools discussed, it is possible for a patient to meet low disease activity or remission cutoffs and still exhibit residual active disease.

PATIENT (PTGA) AND PROVIDER (PRGA) GLOBAL ASSESSMENT OF DISEASE ACTIVITY

Description

Purpose.

The PtGA and PrGA are simple patient-completed or provider-completed visual analog scales (VAS), respectively, measuring the overall way rheumatoid arthritis (RA) affects the patient at a point in time.

Content.

The PtGA includes a statement such as “Considering all of the ways your arthritis has affected you, how do you feel your arthritis is today?” with the best anchor and lowest score on the left side and the worst anchor and highest score on the right side (anchors: very well to very poor). The PrGA includes a statement such as “What is your assessment of the patient's current disease activity?” with the best anchor and lowest score on the left side and the worst anchor and highest score on the right side (anchors: none to extremely active).

Developer/contact information.

E. C. Huskisson, Department of Rheumatology, St. Bartholomew's Hospital Medical College, Charterhouse Square, London ECIM 6BQ, UK (1977).

Versions.

The PtGA and PrGA are classically anchored, unnumbered 10-cm horizontal lines; however, each may be administered as vertical VAS. VAS may be anchored at the ends or open and may have periodic tick marks at specified intervals. Alternatively, a Likert-style VAS may be used. A VAS consisting of 21 circles at 0.5-mm intervals has been shown to be similar to a classic 10-cm line (7).

Number of items.

The PrGA and PtGA are each composed of 1 item. There are no subscales.

Populations.

The VAS was initially designed to be used in measurement of self-assessed pain in RA (8) and has since been used extensively to evaluate overall disease activity. VAS may be used to measure any 1-dimensional aspect of health.

Practical Application

Method of administration.

Patient self-assessment of overall disease activity for PtGA. For PrGA, clinical assessment of overall disease activity performed by the provider.

Administrative burden.

None.

Respondent burden.

PtGA: patient, ∼10 seconds. Provider, ∼10 seconds if using a ruler; use of a VAS consisting of 21 circles may be 5 seconds faster than a VAS requiring use of a ruler (7). Laboratory, not applicable. PrGA: same as PtGA, excluding patient time to complete.

Equipment needed.

A ruler may be required based on the VAS format chosen.

Availability/cost.

There is no cost to use either the PtGA or PrGA.

Scoring.

The PtGA and PrGA each range from 0–100 mm, but are often reported from 0–10 cm.

Score interpretation.

The level of disease activity increases with higher scores.

Method of scoring.

Using a ruler, measure in mm from the left border of the VAS to the point where the patient marked their response on the line. VAS consisting of circles may be scored by visual inspection without use of a ruler. There is no training required to interpret the scores.

Norms available.

Proposed definitions of low disease activity are ≤2.0 (scale 0–10) for the PtGA and ≤1.5 (scale 0–10) for the PrGA (9).

Psychometric Information

Reliability.

Test–retest reliability for the PtGA based on 122 patients tested on an anchored horizontal 10-cm scale with vertical markers every 10 mm demonstrated an interclass correlation coefficient (ICC) of 0.70 (10). Test–retest reliability for the PrGA based on 166 patient encounters on a horizontal 10-cm scale without vertical markers or anchors demonstrated an ICC for test–retest reliability of 0.96 (10). A study of 22 patients demonstrated a κ of 0.58 and an ICC of 0.44 for the PtGA, while the PrGA demonstrated a κ of 0.79 and an ICC of 0.48 (11). For the PtGA. a study of 24 patients demonstrated an ICC of 0.75 (12).

Validity.

Content.

Both the PtGA and PrGA have historically been used to measure disease in RA. The PtGA and PrGA are both American College of Rheumatology (ACR) core set measures for improvement in RA disease activity. The PtGA is incorporated into many composite indices measuring RA disease activity, including the ACR core data set, Disease Activity Score (DAS) and Disease Activity Score with 28-joint counts (DAS28), Clinical Disease Activity Index (CDAI), and Patient-Based Disease Activity Score without erythrocyte sedimentation rate (ESR); the PrGA is also incorporated into composite indices, including the Simplified Disease Activity Index and CDAI. Most patients evaluate their global assessment of RA disease activity higher than providers (13, 14). Interpretation of PrGA and ESR has been shown to demonstrate the least amount of variance among providers as compared to the remaining items in the ACR core set (13).

Concurrent.

Both the PtGA and PrGA have been shown to correlate similarly with radiographic scores as compared to other ACR core set measures, with higher values associated with poorer outcomes; however, association with all ACR core set measures was found to be nonsignificant (P = 0.26) (13). The level of agreement between DAS scores and PrGA was 49%, and between PrGA and ESR was only 17% (15). Significant correlation between initial and longitudinal analysis between PtGA and PrGA has been shown (16).

Convergent.

In clinical trials, the PtGA and PrGA have demonstrated similar change in response to therapy (17, 18). In the measurement of pain, vertical and horizontal VAS have been compared with excellent correlation (0.99) between the 2 scales; however, scores from horizontal scales tend to be slightly lower than those from vertical scales (mean ± SD 10.85 ± 0.63 versus 11.05 ± 0.65) (19). Paper- and computer-based versions of the PtGA are highly correlated (ICC 0.91) (20). In 1 study of 24 patients, the 95% limits of agreement for the smallest detectable difference in PtGA ranged between −41 and 32, suggesting poor reliability as compared to multi-item measures of physical function (12).

Construct.

Duration of morning stiffness does not correlate with PtGA (21) and when compared to the DAS28, the PtGA was 41% more likely to be concordant than discordant (15). Education may affect patient ratings of disease activity (14), comorbid disease increases the PtGA (14), and support groups may decrease PtGA (22). In a study of 24 patients, the PrGA correlated with the PtGA and pain scores (R = 0.2–0.7), with the Health Assessment Questionnaire (HAQ; R = 0.3–0.7), and with ESR (R = 0.2–0.4) (23). In clinical trials, the PtGA and PrGA have demonstrated a similar change in response to therapy (17, 18) and initial PrGA has been shown to correlate with the HAQ 24 months later; however, significant correlation with radiologic outcomes has not been demonstrated (24).

Responsiveness to change.

The smallest detectable difference for the PtGA is large as compared to multi-item measures of physical and psychological function and radiologic measures (12), suggesting poor responsiveness to change. However, PrGA and PtGA, pain scores, and the HAQ have been shown to be more sensitive to change than laboratory measures (23). The PrGA correlates with changes in pain scores, morning stiffness, and ACR functional score, but not with physician-derived swollen joint counts (16). One study found that a worsening in PtGA corresponds to a median DAS28 increase of 0.55 (16.5%) and an improvement in PtGA corresponds to a median DAS28 reduction of −0.82 (16.0%; P < 0.001) (25).

Critical Appraisal of Overall Value to the Rheumatology Community

The PtGA and PrGA are both components of the ACR core set of measures used to assess RA disease activity and are included in many composite indices. The PtGA and PrGA are practical for use in the clinic as little to no training of staff or patients is required and each may be quickly incorporated into the busy clinical setting. Use of a VAS consisting of 21 circles may be slightly faster than a VAS requiring use of a ruler (7) and overcomes the requirement to carry a ruler, which may be cumbersome. The most important concern threatening the validity and reproducibility of both measures is the lack of uniformity regarding the wording of patient and provider instructions and of anchors. As the PtGA and PrGA are overall measures of patient well-being, the effects of comorbid illness (26) and fixed damage, in addition to current RA disease activity, may influence the score. The PtGA may be more a measure of quality of life than of disease activity as it is decreased after participation in support groups (22) and is influenced by patient education (14). Although the PtGA is a component of many composite measures, when used alone it lacks face validity as no provider-derived data or laboratory parameters are included. Additionally, the PtGA has been shown to correlate poorly with changes in the DAS28 (15, 27) and agreement between the PtGA and PrGA is low (κ = 0.17, P < 0.001), indicating that patient perceptions of disease activity may be incongruent with those of providers (28). The PrGA is a component of many composite measures and by itself may encompass all that is known by a provider about a patient's condition, including assessment of joint swelling and pain and interpretation of laboratory values.

DISEASE ACTIVITY SCORE (DAS) AND DISEASE ACTIVITY SCORE WITH 28-JOINT COUNTS (DAS28)

Description

Purpose.

The DAS and DAS28 combine single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. While both the DAS and DAS28 are feasible to use in RA disease activity monitoring, the shorter DAS28 is more feasible for regular clinical use.

Content.

The original DAS includes the number of painful joints calculated by the Ritchie Articular Index (RAI), a 44–swollen joint count (44SJC), erythrocyte sedimentation rate (ESR), and a patient global assessment of disease activity (PtGA) or general health (GH) on a visual analog scale (VAS). The DAS28 includes a 28–swollen joint count (28SJC), 28–tender joint count (28TJC), ESR, and a PtGA or GH assessment on a VAS.

Developer/contact information.

P. L. C. M. van Riel, et al, Department of Rheumatology, University Hospital Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail: P.vanriel@reuma.umcn.nl (DAS: 1990, DAS28: 1995).

Versions.

The DAS28 is analogous to the DAS; however, the DAS28 includes simplified 28-joint counts. The DAS can be used with or without GH assessment and a PtGA may be substituted for GH. A formula to transform the DAS28 into DAS values is available. The DAS and DAS28 may also be calculated using C-reactive protein (CRP) level instead of Westergren ESR. A total of 8 versions are available.

Number of items.

The DAS and DAS28 each have 4 items, or 3 when the GH assessment is omitted. There are no subscales.

Populations.

The target population is patients with RA. It is not formally validated for other rheumatic disorders.

Practical Application

Method of administration.

Clinical assessment of joint counts and patient assessment of disease activity, combined with laboratory evaluation of ESR.

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

Patient, ∼10 seconds; provider, ∼5–8 minutes for DAS and 3–5 minutes for DAS28; laboratory, 1 hour waiting time for ESR, waiting time for CRP level varies by laboratory.

Equipment needed.

A ruler may be required based on the VAS format chosen. A laboratory is needed for ESR or CRP determination. A programmed calculator or computer is needed.

Availability/cost.

The DAS and DAS28 may be used free of charge and are available online at http://www.das-score.nl.

Scoring.

The RAI ranges from 0–78. The 44SJC ranges from 0–44. The 28SJC and 28TJC each range from 0–28. The GH/PtGA VAS ranges from 0–100. The ESR generally ranges from 0–100 mm/hour. Normal CRP level varies by laboratory and is expressed in mg/liter, with use of a CRP test with a lower detection level of 1.0 mg/liter advised. The ranges of the DAS and DAS28 are 0–10 and 0–9.4, respectively.

Scores interpretation.

For all DAS versions, the level of disease activity can be interpreted as remission (DAS <1.6), low (1.6 ≤ DAS < 2.4), moderate (2.4 ≤ DAS ≤ 3.7), or high (DAS >3.7) (29). For all DAS28 versions, the level of disease activity can be interpreted as remission (DAS28 <2.6), low (2.6 ≤ DAS28 < 3.2), moderate (3.2 ≤ DAS28 ≤ 5.1), or high (DAS28 >5.1). Alternative cutoff values for the DAS28 have been proposed that include a more stringent remission of ≤2.4 (30).

A change of 1.2 (2 times the measurement error) in either the DAS or DAS28 is considered a significant change (29). The European League Against Rheumatism (EULAR) response criteria classify patients as good, moderate, or nonresponders, using both the change in DAS and level of DAS reached. Good response is defined as improvement in DAS >1.2 and followup DAS ≤2.4; nonresponders have improvement in DAS ≤0.6 or improvement >0.6 but ≤1.2 and followup DAS >3.7; all others are classified as moderate responders (29).

Method of scoring.

The following formulas are used: 1) DAS-4 (4 variables) = 0.54 × √(RAI) + 0.065 × (44SJC) + 0.33 × ln(ESR) + 0.0072 × GH; 2) DAS-3 (3 variables) = 0.54 × √(RAI) + 0.065 × (44SJC) + 0.33 × ln(ESR) + 0.22; 3) DAS-4(CRP) = 0.54 × √(RAI) + 0.065 × (44SJC) + 0.17 × ln(CRP + 1) + 0.0072 × GH + 0.45; 4) DAS-3(CRP) = 0.54 × √(RAI) + 0.065 × (44SJC) + 0.17 × ln(CRP + 1) + 0.65; 5) DAS28(4) = 0.56 × √(28TJC) + 0.28 × √(28SJC) + 0.70 × ln(ESR) + 0.014 × GH; 6) DAS28(3) = [0.56 × √(28TJC) + 0.28 × √(28SJC) + 0.70 × ln(ESR)] × 1.08 + 0.16; 7) DAS28-CRP(4) = 0.56 × √(28TJC) + 0.28 × √(28SJC) + 0.36 × ln(CRP + 1) + 0.014 × GH + 0.96; and 8) DAS28-CRP(3) = [0.56 × √(28TJC) + 0.28 × √(28SJC) + 0.36 × ln(CRP + 1)] × 1.10 + 1.15.

The DAS and DAS28 are not directly comparable; however, the following formula may be used to transform scores: DAS28 = (1.072 × DAS) + 0.938. There is no training required to interpret the scores.

Norms available.

Reference values are available and are useful for the interpretation of the DAS from individual patients (see above).

Psychometric Information

Reliability.

Test–retest.

Test–retest reliability for the DAS ranges from 0.8–0.89 (29, 31). Test–retest reliability for the DAS with 3 variables was 0.89 (31). When repeated testing 1 week apart was performed, the DAS28 demonstrated a correlation of 0.85 and an interclass correlation coefficient of 0.85 (P < 0.01) (32).

Reproducibility.

The DAS was designed based on statistical analysis of factors governing clinical judgment in usual rheumatologic care in a cohort of early arthritis patients (33); subsequently, principal component analysis has found no influence of disease duration on the structure of the DAS and DAS28 scores (34).

Validity.

Content.

The DAS and DAS28 include variables from the American College of Rheumatology (ACR) core set of measures used to assess outcomes in RA. No measures of disability or joint damage are included. Due to inclusion of ESR, which is influenced by age, the DAS and DAS28 may underestimate remission in the elderly (35).

Criterion.

Physician judgment of disease activity level was used as an external standard for DAS and DAS28 development (33, 34). The DAS and DAS28 have been shown to discriminate active from inactive RA disease activity (36), with predictive abilities of 0.93 for the DAS (37) and 0.88 for the DAS28 (38). DAS28 and DAS 28-CRP scores have been shown to discriminate between ACR response categories (39), with good agreement between DAS28-CRP and EULAR response rates at 6 months (κ = 0.80–0.82) (40, 41). DAS28 scores also agree with Clinical Disease Activity Index (CDAI) values (κ = 0.70) (39). Despite nonsignificance in classification accuracy (42), the frequency of classifying patients in DAS28 remission is higher than for the Simplified Disease Activity Index (SDAI), CDAI (43, 44), and DAS (45), and is variable compared to modified ACR criteria (46, 47). Misclassification of remission has been shown high as 11.2% when comparing the DAS28-CRP to the DAS28 (48).

Convergent.

The DAS and DAS28 are highly correlated (R = 0.97) (34), as are DAS28 and DAS28-CRP values (R = 0.95); however, correlation between the DAS28 and DAS28-CRP is relatively weak at lower values (48), with DAS28-CRP scores generally lower than DAS28 scores (40). The mean of the differences between the DAS28 ESR- and CRP-based scores is 0.34 (95% limits of agreement −0.58, 1.26) (40). The DAS28 has been shown to explain 9.8% of the total variance in activity restriction (49) and significantly contributes to patient-perceived work disability risk independently from Health Assessment Questionnaire (HAQ) scores (50); however, the DAS28 demonstrates variable levels of correlation with HAQ scores (R = 0.32–0.68, P < 0.05), which may be related to disease duration (39, 51, 52). DAS (<1.6) and DAS28 (<2.6) remission may be discordant in up to 15% of patients with 96% of discordant pairs meeting DAS28 but not DAS remission and 4% meeting DAS but not DAS28 remission, suggesting the DAS28 may potentially overestimate remission due to residual disease activity in joints not included in 28-joint counts (45). Approximately 70% of patients classified as being in remission by the DAS28 (DAS28 <2.4) have no swollen joints, while 85% of patients classified as being in remission by the SDAI have no swollen joints (4). The DAS28 correlates highly with provider (R = >0.94) and to a lesser degree with patient assessment of disease activity (R = >0.60) (53). The DAS and DAS28 are variably correlated with power Doppler ultrasound measurement of joint inflammation (54, 55).

Predictive.

The DAS and DAS28 do not correlate strongly with joint damage at a single point in time (56, 57). There is a linear correlation between the DAS28 and radiographic progression (R = 0.58, P < 0.0001) (39) with patients who spend >50% of their time in DAS28 remission having less radiographic progression than those not in remission at least 50% of the time (4). EULAR good responders have demonstrated less radiographic progression at up to 18 months (P = 0.0001) (58) with EULAR and ACR response criteria, demonstrating equivalent radiographic progression (59). A study evaluating progression of elbow arthritis demonstrated that a cutoff of 3.15 for mean DAS28-CRP(3) at 0–2 years of disease diagnosis predicted elbow deterioration at 10 years (60).

Construct.

Increases in the DAS and DAS28 have been shown to correlate with worsening function as measured by the HAQ and Short Form 36 physical functioning scale (51, 56, 61). The DAS28 and DAS28-CRP have both been shown to have similar mean changes in HAQ scores (−0.86 and −0.80 for good responders, −0.46 and −0.41 for moderate responders, and 0.05 and 0.05 for nonresponders, respectively) (41).

Responsiveness to change.

In a trial comparing sulfasalazine with a combination of methotrexate, sulfasalazine, and prednisolone, all composite indices were more responsive than single ACR core set measures. The ACR 20% response criteria were more responsive; the standardized response mean (SRM; mean change divided by SD of change) of the DAS was approximately half as large (2). The SRM for the DAS28 has been shown to be slightly higher than that of the DAS (38).

Critical Appraisal of Overall Value to the Rheumatology Community

Both the DAS and DAS28 have been extensively validated, are endorsed by the ACR and EULAR for RA disease activity measurement in clinical trials (62), and are often considered the “gold standard” by which to measure RA disease activity, with the majority of newer disease activity monitoring tools based on, or compared to, the DAS28. Notably, remission classification for the DAS28 is less conservative than for several other measurement tools, including the DAS, SDAI, and CDAI (40, 43, 45). The DAS and DAS28 were designed based on statistical analysis of factors governing clinical judgment in usual rheumatologic care in a cohort of early arthritis patients (33, 34) with responsiveness of the composite measure repeatedly demonstrated (59). Provider-derived measurement of RA disease activity combined with a laboratory value and a small contribution from patient-derived data gives good face validity. Disadvantages of the DAS and DAS28 in clinical practice are the need for a blood sample, time needed for providers to perform joint counts, complicated mathematical calculation of the composite score, and potential confusion from choosing among the multiple formulas available. Importantly, treating to a target DAS level <2.4 (DAS28 <3.6) has been shown to improve RA outcomes (63, 64). Neither the RAI nor the 28TJC and 28SJC can be considered superior (65) and reduced and expanded joint counts have shown similar sensitivity (66), although concern remains that 28-joint counts may miss important disease activity present in uncounted joints (45). While clinicians tend to give swollen joint counts more importance when making treatment decisions (67), the DAS28 formula weighs tender joints more heavily than swollen joints. Additionally, ESR has been shown to contribute 15% of the information in the DAS28 (30) and remission may be underestimated in high ESR states. Conversely, in low ESR states it is possible to be in DAS or DAS28 remission and still have large numbers of swollen joints (68, 69). Caution should be employed when using the DAS28-CRP as a substitute for the DAS28, as it produces lower scores (40, 41, 48) and may underestimate the level of active disease. Limited validation of the DAS28-3 has been performed; however, proposed remission values have been shown to correlate highly with disease activity (R = 0.95, P < 0.0001) (70), making the DAS28-3 of use when PtGA data are lacking.

SIMPLIFIED DISEASE ACTIVITY INDEX (SDAI)

Description

Purpose.

The SDAI combines single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. It is feasible to use for monitoring of RA disease activity in daily clinical practice.

Content.

The SDAI includes a 28–swollen joint count (28SJC), 28–tender joint count (28TJC), patient global assessment of disease activity (PtGA) on a 10-cm visual analog scale (VAS), provider global assessment of disease activity (PrGA) on a 10-cm VAS, and C-reactive protein (CRP) level in mg/dl.

Developer/contact information.

Josef Smolen, MD, et al, Department of Rheumatology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. E-mail: Josef.smolen@wienkav.at (2003).

Versions.

The SDAI is analogous to the Clinical Disease Activity Index (CDAI); however, the SDAI includes laboratory measurement of CRP level.

Number of items.

The SDAI has 5 items. There are no subscales.

Populations.

The target population is patients with RA. The SDAI is not validated for other rheumatic disorders.

Practical Application

Method of administration.

Clinical assessment of joint counts and PrGA combined with self-administered patient assessment of disease activity.

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

Patient, ∼10 seconds; provider, ∼2 minutes; laboratory, waiting time for CRP level varies by laboratory.

Equipment needed.

A ruler may be required based on the VAS format chosen. A laboratory is needed for CRP level determination. For calculation of the SDAI, a calculator may be used; however, it is not required.

Availability/cost.

The SDAI may be used free of charge; no specialized form is needed.

Scoring.

The 28SJC and 28TJC each range from 0–28. The PtGA and PrGA each range from 0–10. Normal CRP level varies by laboratory and is expressed in mg/dl. The lower range of the SDAI is 0, with the upper end resting on the upper limit of CRP level, often defined as 10 mg/dl, leading to an upper limit of 86.

Score interpretation.

The level of disease activity can be interpreted as remission (SDAI ≤3.3), low (3.3 < SDAI ≤ 11), moderate (11 < SDAI ≤ 26), or high (SDAI >26) (30). The smallest detectable difference for the SDAI is 8.26 (32). An SDAI change of 16 corresponds to a change of 1.2 for the Disease Activity Score with 28-joint counts (DAS28) and an SDAI change of 20.7 corresponds to a 0.22 change in Health Assessment Questionnaire (HAQ) score, both indicating clinically significant change (29, 71, 72).

Method of scoring.

The SDAI is calculated by adding the 5 items together: SDAI = 28SJC + 28TJC + PrGA + PtGA + CRP. There is no training required to interpret the scores.

Norms available.

Reference values are available, and are useful for the interpretation of the disease activity scores from individual patients (see above).

Psychometric Information

Reliability.

Test–retest.

When repeated testing 1 week apart was performed, the SDAI demonstrated a correlation of 0.87 and an interclass correlation coefficient of 0.88 (P < 0.01) (32). The components of the SDAI are individually accepted as reliable in RA assessment.

Reproducibility.

The SDAI was originally validated with additional patients within the original study (71) and later validated in multiple additional data sets (30). Reproducibility of the SDAI was evaluated in a cross-sectional inception cohort of newly diagnosed RA patients seen every 3 months for 1 year (n = 91) and an observational routine care cohort (n = 279) with significant correlations at the P < 0.001 level as compared to the DAS28, CDAI, and HAQ (30).

Validity.

Content.

The SDAI includes variables from the American College of Rheumatology (ACR) core set of measures used to assess outcomes in RA. No measures of disability or joint damage are included. The SDAI was evaluated using a survey of patient profiles among 21 rheumatologists that showed excellent agreement with the physician's ratings of the patients' disease activity (71).

Criterion.

The SDAI demonstrates linear correlation with the DAS28 (R = 0.80–0.92, P < 0.0001) and is highly related to patient-reported pain (R = 0.66, P < 0.001) (73). Major improvement in the SDAI as defined by a decrease of at least 22 at 12 months corresponds to a mean increase of total Sharp score of 1.1, identical to Sharp score progression in good Disease Activity Score (DAS) responders. Larsen scores confirm results of Sharp scores (71). The SDAI is 90% sensitive and 86% specific for prediction of clinicians' decision to change disease-modifying antirheumatic drug therapy and surpasses predictive ability of the DAS28 due to weighting of swollen joint scores (74). Approximately 85% of patients classified as being in remission by the SDAI have no swollen joints; 70% of patients classified as being in remission by the DAS28 have no swollen joints (4). SDAI scores are variably correlated with joint inflammation by power Doppler ultrasound (54, 55).

Convergent.

Median SDAI scores (11.6, range 0.07–46.60) are slightly higher (P < 0.001) than median CDAI scores (10.7, range 0–42.10) and demonstrate sex differential with median SDAI scores of 12.2 (range 0.07–46.60; P < 0.001) in women and 8.0 (range 0.10–35.20; P < 0.001) in men (73).

Predictive.

One study showed that over 3 years, patients who spend the majority of time in SDAI remission do not progress radiographically as compared to those who spent ≤50% time in remission (4). The SDAI was sensitive in discriminating between different ACR response categories and HAQ change scores (P < 0.0001) (39).

Construct.

From baseline to SDAI remission, HAQ score improvement differs by degree of joint damage with ∼25% of patients with moderate to severe joint damage having <50% improvement in the HAQ as compared to patients with little joint damage having >80% improvement in the HAQ (75).

Responsiveness to change.

As compared to the DAS28, HAQ, and ACR 20% response criteria, the SDAI demonstrates a fairly consistent and proportional change with changes in the SDAI increasing slightly with time (71). Change in the SDAI exhibits a linear relationship with change in the HAQ (R = 0.56–0.57, P < 0.0001) and the modified HAQ (R = 0.48, P < 0.0001) (71).

Critical Appraisal of Overall Value to the Rheumatology Community

The SDAI was developed to simplify complicated calculations and overcome the shortcomings of the DAS. The SDAI constitutes a simple numerical addition of individual measures on their original scale, overcoming problems of transformations and weighting used in other composite indices. The SDAI is endorsed by the ACR and European League Against Rheumatism (EULAR) for RA disease activity measurement in clinical trials and by EULAR for patient monitoring (6, 62). There is no need to use a calculator or computer for calculations. The inclusion of both patient- and provider-derived data as well as a laboratory marker of inflammation gives face validity to the SDAI. The use of CRP in mg/dl, rather than erythrocyte sedimentation rate, is advantageous as it does not overweigh the laboratory variable in the final score. Pain assessment was not included in the SDAI as it is also reflected in the patient global estimate of disease activity. The HAQ disability index was not included in the SDAI as it is not routinely done in clinical practice by many providers (71). Additionally, loss of function as measured by the HAQ or other tools may be irreversible, with some considering addition of a functional measure in a composite index of disease activity confusing despite noting the importance of functional information in patient care (4). Although sex appears to have a small effect on the SDAI score (73), different disease activity cutoffs for male and female patients are not established.

CLINICAL DISEASE ACTIVITY INDEX (CDAI)

Description

Purpose.

The CDAI combines single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. It is feasible to use the CDAI for monitoring of RA disease activity in daily clinical practice, and lack of a laboratory value may facilitate more routine evaluation of RA disease activity.

Content.

The CDAI includes a 28–swollen joint count (28SJC), 28–tender joint count (28TJC), patient global assessment of disease activity (PtGA) on a 10-cm visual analog scale (VAS), and provider global assessment of disease activity (PrGA) on a 10-cm VAS.

Developer/contact information.

Daniel Aletaha, MD, et al, Department of Rheumatology, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. E-mail: aletaha@mail.nih.gov (2005).

Versions.

The CDAI is analogous to the Simplified Disease Activity Index (SDAI); however, the CDAI excludes laboratory measurement of C-reactive protein (CRP) level.

Number of items.

The CDAI has 4 items. There are no subscales.

Populations.

The target population is patients with RA. It is not validated for other rheumatic disorders.

Practical Application

Method of administration.

Clinical assessment of joint counts and PrGA combined with self-administered patient assessment of disease activity.

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

Patient, ∼10 seconds; provider, <2 minutes; laboratory, not applicable.

Equipment needed.

A ruler may be required based on the VAS format chosen. For calculation of the CDAI, a calculator may be used; however, it is not required.

Availability/cost.

The CDAI is available free of charge; no specialized form is needed.

Scoring.

The 28SJC and 28TJC each range from 0–28. The PtGA and PrGA each range from 0–10. The range of the CDAI is 0–76.

Score interpretation.

The level of disease activity can be interpreted as remission (CDAI ≤2.8), low (2.8 < CDAI ≤ 10), moderate (10 < CDAI ≤ 22), or high (CDAI >22) (30), although others have reported differing cutoffs (76). The smallest detectable difference for the CDAI is 8.05 (32).

Method of scoring.

The CDAI is calculated by adding the 4 items together: CDAI = 28SJC + 28TJC + PrGA + PtGA. There is no training required to interpret the scores.

Norms available.

Reference values are available, and are useful for the interpretation of the disease activity scores from individual patients (see above).

Psychometric Information

Reliability.

Test–retest.

When repeated testing 1 week apart was performed, the CDAI demonstrated a correlation of 0.89 and an interclass correlation coefficient of 0.89 (P < 0.01) (32). The components of the CDAI are individually accepted as reliable in RA assessment.

Reproducibility.

Reproducibility of the CDAI was evaluated in a cross-sectional inception cohort of RA patients seen every 3 months for 1 year (n = 91) and observational routine care cohorts (n = 279) with significant correlations (P < 0.001) as compared to the Disease Activity Score with 28-joint counts (DAS28), CDAI, and Health Assessment Questionnaire (HAQ) (30). The CDAI was sensitive in discriminating between different American College of Rheumatology (ACR) response categories and HAQ change scores (P < 0.0001) (39).

Validity.

Content.

The CDAI includes variables from the ACR core set of measures used to assess outcomes in RA. No measures of disability or joint damage are included. The CDAI has been shown to correlate closely with the SDAI, with only 5% of the variance in the SDAI explained by CRP level (39). The CDAI is highly related to patient-reported pain (R = 0.67, P < 0.001) (73).

Criterion.

In validation studies, the CDAI showed a correlation coefficient as compared to the DAS28 of 0.89 (P < 0.001) (39). CDAI scores have been shown to correlate with ultrasound-derived swollen joint scores (R = 0.6, P < 0.006) (55).

Convergent.

Median CDAI scores (10.7, range 0–42.10) are slightly lower (P < 0.001) than median SDAI scores (11.6, range 0.07–46.60) and demonstrate sex differential with median CDAI scores of 11.3 (range 0.00–42.10; P < 0.001) in women and 7.1 (range 0.00–32.00; P < 0.001) in men (73).

Predictive.

CDAI scores correlate with radiographic change as measured by Larsen scores (R = 0.59, P < 0.0001) (39).

Responsiveness to change.

The CDAI demonstrates a linear relationship with the HAQ/modified HAQ (R = 0.47–0.56, P < 0.0001) (71) and demonstrates the ability to discriminate degrees of ACR response (39).

Critical Appraisal of Overall Value to the Rheumatology Community

Based on the SDAI, the CDAI was developed as a simple calculation of disease activity for use in the clinic at point of care. The CDAI constitutes a simple numerical addition of individual measures on their original scale, overcoming problems of transformations and weighting used in other composite indices. The CDAI is endorsed by the ACR and European League Against Rheumatism (EULAR) for RA disease activity measurement in clinical trials and by EULAR for patient monitoring (6, 62). There is no need to use a calculator or computer for calculations. Like the SDAI, no functional or quality of life index was included in the CDAI. Like the SDAI, one drawback of the CDAI is not all physicians perform detailed joint examinations when assessing RA disease activity. Additionally, the CDAI does not include any laboratory measurement and therefore all variables are easily available at point of care in the clinical setting, which can in turn produce more consistency in timing and completeness of disease measurement. While exclusion of laboratory measurements decreases face and content validity of the composite measure, similar correlations to the DAS28 and HAQ have been demonstrated as well as discrimination between ACR response categories with and without the addition of CRP level to the remaining variables (39). Additionally, although acute-phase reactants correlate with long-term outcomes in RA, they have not been shown to contribute sufficient information in other composite scores to change judgment of disease activity (39). Although sex appears to have a small effect on the CDAI score (39), different disease activity cutoffs for male and female patients are not established.

PATIENT ACTIVITY SCORE (PAS) AND PATIENT ACTIVITY SCORE-II (PASII)

Description

Purpose.

The PAS and PASII combine single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. It is feasible to use the PAS/PASII for monitoring of RA disease activity in daily clinical practice as the lack of a laboratory value, or provider-derived data may facilitate more routine evaluation of RA disease activity when time constraints are considered.

Content.

The PAS and PASII contain only patient-derived data and include a patient assessment of pain on a 10-cm visual analog scale (VAS), patient global assessment of disease activity (PtGA) on a 10-cm VAS, and the Health Assessment Questionnaire (HAQ) for the PAS or the Health Assessment Questionnaire-II (HAQII) for the PASII.

Developer/contact information.

Frederick Wolfe, National Data Bank for Rheumatic Diseases, Arthritis Research Center Foundation, 1035 North Emporia, Suite 288, Wichita, KS 67214. E-mail: fwolfe@arthritis-research.org (2005).

Versions.

The PAS is analogous to the PASII and Routine Assessment of Patient Index Data 3 (RAPID3); however, the PASII substitutes the HAQII for the HAQ and RAPID3 substitutes the Multidimensional Health Assessment Questionnaire (MDHAQ) for the HAQ. Spanish versions of the PAS and PASII are available.

Number of items.

The PAS and PASII have 3 items each. There are no subscales.

Populations.

The target population is patients with RA. The PAS has not been validated for other rheumatic disorders. The PASII has also been used in other rheumatic diseases, including ankylosing spondylitis, psoriatic arthritis, gout, osteoarthritis, fibromyalgia, systemic lupus erythematosus, systemic sclerosis, polymyalgia rheumatica, inflammatory myositis, Sjögren's syndrome, Behçet's disease, and other inflammatory and noninflammatory rheumatic diseases (77).

Practical Application

Method of administration.

Patient-administered questionnaire.

Administrative burden.

None.

Time to complete.

PAS: patient, <3.5 minutes; provider, <1 minute; laboratory, not applicable. PASII: patient, <1.5 minutes; provider, <30 seconds; laboratory, not applicable.

Equipment needed.

A ruler may be required based on the VAS format chosen. For calculation of the PAS and PASII, a calculator may be used; however, it is not required.

Availability/cost.

The PAS and PASII may be used free of charge and components are available online at http://www.arthritis-research.org/research/pas.

Scoring.

The HAQ and HAQII range from 0–3. The patient assessment of pain and PtGA VAS both range from 0–10. The range of both the PAS and PASII is 0–10.

Score interpretation.

For the PAS, categories of disease activity may be interpreted as remission (PAS ≤1.25) and low (PAS ≤1.75) (78). For the PASII, categories of disease activity may be defined as remission (PASII ≤1.25) and low (PASII ≤2.2) (43). Categories of self-reported disease severity for the PAS and PASII may be defined as remission = ≤0.25, low = ≤3.7, moderate = <8.0, and high = ≥8.0. (Michaud K, et al: unpublished observations).

Method of scoring.

The PAS is calculated by multiplying the HAQ by 3.33 and then dividing the sum of the VAS pain, PtGA, and HAQ by 3. For the PASII, the HAQ is replaced by the HAQII. There is no training required to interpret the scores.

Norms available.

Specific reference values are available (see above); however, disease activity cutoffs for the PAS and PASII have been variably reported.

Psychometric Information

Reliability.

Test–retest reliability studies have not been done for the PAS or PASII.

Validity.

Content.

The PAS and PASII contain patient-derived variables from the American College of Rheumatology (ACR) core set of measures used to assess outcomes in RA. No provider-derived data are included in the composite scores.

Criterion.

The κ value for remission between the PASII and Disease Activity Score with 28-joint counts (DAS28) is 0.25, for the PASII and provider global assessment of disease activity is 0.29, and for the PASII and Clinical Disease Activity Index is 0.40 (43). The PAS has not been evaluated against the Disease Activity Score (DAS), DAS28, or ACR response criteria. An increase of 1 point in the PASII had an odds ratio of 1.19 (95% confidence interval 1.07–1.33) for predicting active from inactive disease independent of specific rheumatic disease diagnosis (77).

Concurrent.

The PAS and PASII are equally associated with the Short Form 36 (SF-36) composite scale and the EuroQol (Kendall's τ-a = 0.59). The SF-36 physical and mental component subscales and the Rheumatoid Arthritis Disease Activity Index are more strongly associated with the PAS/PASII as compared to the EuroQol. As compared to the HAQ and HAQII, the PAS and PASII satisfactorily predict mortality, performing only marginally less well (79).

Responsiveness to change.

Studies to evaluate for responsiveness to change have not been done for the PAS or PASII. Both the PAS and PASII are comprised of ACR core set measures that have documented responsiveness to change; however, pain scores have been shown to be relatively stable over time in established RA (80).

Critical Appraisal of Overall Value to the Rheumatology Community

The PAS and PASII are simple to use composite scores that use patient-derived data without provider effort, which makes the feasibility of the measure very good for use in daily clinical care. While the PAS and PASII are composed of the 3 patient-derived ACR core set data elements, the measures have not been compared directly to measures such as the ACR response criteria or the DAS in clinical trials and may not be suitable for use in that setting. As the components of the PAS are included in the majority of observational and clinical studies, it has been suggested that the PAS could be used to compare data between such studies (79). Although the HAQII is not widely used at this time, it possesses superior psychometric properties to the MHAQ and MDHAQ, and a formula exists to convert from the HAQII to the HAQ (81, 82). Other than the choice of HAQ version, the PAS is identical to both the PASII and RAPID3. An advantage of the PASII as compared to the PAS is use of the shorter HAQII, which is therefore easier for patients to complete. Time to complete the PASII is identical to RAPID3; however, the HAQII is impacted less by skipped questions and has a lesser floor effect as compared to the MDHAQ (82).

ROUTINE ASSESSMENT OF PATIENT INDEX DATA (RAPID)

Description

Purpose.

The RAPID2 and RAPID3 are brief self-administered questionnaires of disease symptoms for patients with rheumatoid arthritis (RA). RAPID4 patient joint count (RAPID4PTJC) adds a self-reported joint count to RAPID3. RAPID4 provider joint count (RAPID4MDJC) and RAPID5 combine self-administered questionnaires of disease symptoms with a provider-derived joint count or provider global assessment of disease activity (PrGA), respectively.

Content.

The RAPID scores include combinations of the following 2–5 items of the following 6 items: the Multidimensional Health Assessment Questionnaire (MDHAQ), a pain visual analog scale (VAS), patient global assessment of disease activity (PtGA) on a 10-cm VAS, PrGA on a 10-cm VAS, swollen joint count (SJC), and tender joint count (TJC).

Developer/contact information.

Theodore Pincus, MD, Director of Outcomes Research, Division of Rheumatology, Department of Medicine, New York University Hospital for Joint Diseases, New York University Medical Center, 301 East 17th Street, Room 1608, New York, New York 10003. Email: tedpincus@gmail.com (1999).

Versions.

There are 5 RAPID score versions: RAPID2 includes the PtGA and the PrGA; RAPID3 includes the MDHAQ, a pain VAS, and the PtGA; RAPID4TJC includes the MDHAQ, a pain VAS, PtGA, and the Rheumatoid Arthritis Disease Activity Index (RADAI) self-reported TJC; RAPID4MDJC includes the MDHAQ, a pain VAS, PtGA, and provider-derived 28TJC and 26SJC (excludes shoulders); and RAPID5 includes the MDHAQ, a pain VAS, PtGA, PrGA, and the RADAI self-reported TJC (83, 84). A Spanish version is available.

Number of items.

The RAPID scores have 2–5 items, corresponding to the number after RAPID (i.e., RAPID3 has 3 items). There are no subscales.

Populations.

The target population is patients with RA. RAPID3 has been used in systemic lupus erythematosus, spondylarthropathies, vasculitis, psoriatic arthritis, gout, scleroderma, osteoarthritis, fibromyalgia, familial Mediterranean fever, and Behçet's disease (85).

Practical Application

Method of administration.

For RAPID score versions requiring a joint count, either provider joint counts or patient self-reported RADAI joint counts may be used to calculate RAPID scores (84).

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

RAPID2: patient, ∼20 seconds; provider, <20 seconds; laboratory, not applicable (N/A). RAPID3: patient, ∼1.5 minutes; provider, <30 seconds; laboratory, N/A. RAPID4PTJC: patient, 5–10 minutes; provider, <1 minute; laboratory: N/A. RAPID4MDJC: patient, <1.5 minutes; provider, ∼2 minutes; laboratory, N/A. RAPID5: patient, 5–10 minutes if patient joint count and <5 minutes if provider joint count; provider, <1 minute if patient joint count and <2.5 minutes if provider joint count; laboratory, N/A (86–88).

Equipment needed.

A ruler may be required based on the VAS format chosen. For calculation of RAPID scores a scoring template or calculator may be used, but is not required.

Availability/cost.

The RAPID scores are all available free of charge. The RAPID3 is available online at http://mdhaq.org/Public/Questionnaires.aspx.

Scoring.

The RADAI self-report joint count ranges from 0–48. Provider tender joint scores range from 0–28. Provider swollen joint scores range from 0–26. The MDHAQ ranges from 0–3. Pain, PtGA, and PrGA are scored from 0–10, each on a separate 10-cm horizontal VAS, which may be substituted for a 21-circle VAS. The range of each of the RAPID scores is 0–10. Scores can be calculated by hand, by calculator, or by use of a scoring template.

Score interpretation.

Recommended cutoffs for disease activity are not available for RAPID2 scores. RAPID3, RAPID4PTJC, RAPID4MDJC, and RAPID5 scores range from 0–1.0 = remission, 1.1–2.0 = low, 2.1–4.0 = moderate, and 4.1–10 = high (89). Alternatively, if the RAPID3 is scored on a 0–30 scale, the following cutoffs for disease activity may be used: remission = 0–3.0, low = 3.1–6.0, moderate = 6.1–12.0, and high = 12.1–30 (86). The smallest detectable difference for RAPID3 is 1.48 (32).

Method of scoring.

All raw scores may be converted to a range of 0–10. The MDHAQ score (0–3) is multiplied by 3.33. Pain, PtGA, and PrGA are scored from 0–10, each on a separate 10-cm VAS. Tender and swollen joint scores are converted to a 0–10 scale based on simple division (i.e., for 28-joint count, divide by 2.8). After the component scores are standardization to 0–10 scales, the individual items in the desired composite are added together and then divided by the number of items in the composite to give an adjusted final score (i.e., RAPID3 raw score of 0–30 is divided by 3 to give the final adjusted score ranging 0–10) (84). There is no training required to interpret the scores.

Norms available.

See above.

Psychometric Information

Reliability.

When repeated testing 1 week apart was performed, RAPID3 demonstrated a correlation of 0.88 and an interclass correlation coefficient of 0.90 (P < 0.01) (32). Test–retest reliability has not been evaluated for the other RAPID scores.

Validity.

Content.

The RAPID scores are all composed of American College of Rheumatology (ACR) core set measures used to assess the efficacy of disease-modifying antirheumatic drugs. None of the RAPID scores includes acute-phase reactant values.

Criterion.

The RAPID3, RAPID4MDJC, RAPID4PTJC, and RAPID5 scores demonstrate Spearman's correlation coefficients as compared to the Disease Activity Score with 28-joint counts (DAS28) between 0.69 and 0.73, with the RAPID4MDJC demonstrating the highest value. Spearman's correlation coefficients for each of the RAPID scores (excluding RAPID2) as compared to the Clinical Disease Activity Index (CDAI) ranged from 0.74–0.83, with the RAPID4MDJC again demonstrating the highest value. Correlation coefficients between each of the RAPID scores (excluding RAPID2) were between 0.98 and 0.99. An additional study demonstrated Spearman's correlation coefficients between 0.36 and 0.61 for RAPID3 as compared to the DAS28 and between 0.54 and 0.77 as compared to the CDAI (86).

Concurrent.

When comparing RAPID3 to ACR 20% improvement criteria and the DAS in a randomized controlled trial, the κ ranged between 0.62 and 0.64, indicating similar sensitivity in identification of active treatment from placebo (90). In another study, κ values ranged from 0.22–0.37 for CDAI versus RAPID3 and from 0.12–0.20 for DAS28 versus CDAI, indicating a fair agreement of RAPID3 to both of the indices (86).

Construct.

Addition of TJC or SJCs and/or physician estimate of global status does not add to the capacity of RAPID3 to distinguish active from control treatments, and all RAPID scores perform similarly in this function (84).

Responsiveness to change.

Studies to evaluate for responsiveness to change have not been done for the RAPID scores. Each of the RAPID scores is comprised of combinations of ACR core set measures that have well-documented responsiveness to change; however, pain scores have been shown to be relatively stable over time in established RA (80).

Critical Appraisal of Overall Value to the Rheumatology Community

The RAPID scores are feasible in the clinical setting and are based on ACR criteria. The flexibility to choose which of the RAPID scores used is advantageous. The RAPID3 score is the most frequently used and best-validated measure among the various RAPID scores. RAPID4 scores have been presented in 2 different forms with use of TJCs or both TJCs and SJCs and as such, standardization is lacking. Additionally, RAPID4 and RAPID5 scores have been reported in the literature using either provider- or patient-derived joint counts ranging from 28–78 joints. This lack of standardization makes these scores more difficult to compare between users. Furthermore, both provider and patient joint counts are poorly reproducible (91–93) and patient-reported TJCs correlate only moderately with physician joint counts, with SJCs demonstrating lower levels of correlation (94). More recently, Pincus et al have advocated using the RADAI self-reported joint count (95), which would partially eliminate issues of variability among joint assessment. While RAPID2 scores are the easiest to perform in a busy clinical setting, they also provide the least information as compared to the other RAPID scores. Patient-derived measures without acute-phase reactant values have been shown to be reliable and sensitive (96); however, they may be influenced by patient education level (97). While the use of the MDHAQ is shorter and easier to administer than the original HAQ, it has a greater floor effect and missing 1 or more items has more impact on the score than the newer HAQ-II (81, 82), which is a component of the otherwise identical Patient Activity Score-II. Additionally, despite concerns that patient questionnaires may reflect irreversible joint damage, RAPID scores have demonstrated similar efficiency to joint counts in differentiating active from inactive disease (86).

RHEUMATOID ARTHRITIS DISEASE ACTIVITY INDEX (RADAI) AND RHEUMATOID ARTHRITIS DISEASE ACTIVITY INDEX-5 (RADAI-5)

Description

Purpose.

The RADAI and RADAI-5 are patient-assessed measures for disease activity in rheumatoid arthritis (RA). Both the RADAI and RADAI-5 may complement or replace the physician's assessment of disease activity in health services or epidemiologic research, and may be used for patient management.

Content.

The RADAI and RADAI-5 are both 5-item questionnaires. The items ask about global disease activity in the last 6 months, current disease activity with respect to joint tenderness and swelling, arthritis pain, and duration of morning stiffness. The RADAI also includes tender joints rated on a joint list and the RADAI-5 asks about general health.

Developer/contact information.

RADAI: Gerold Stucki, Department of Physical Medicine and Rehabilitation, University Hospital Munich, Marchionistrasse 15, D-81377 Munich, Germany. E-mail: gerold.stucki@phys.med.uin-muenchen.de (1995). RADAI-5: Burkhard F. Leeb, Karl Landsteiner Institute for Clinical Rheumatology, Landstrasse 18, A-2000 Stockerau, Austria. E-mail: burkhard.leeb@stockerau.lknoe.at (2008).

Versions.

The RADAI-5 is analogous to the RADAI with replacement of joint counts by the patient's general health assessment and slight change of question format. The RADAI is also available in German, French, Italian, and Dutch. The RADAI-5 is also available in German.

Number of items.

The RADAI and RADAI-5 each contain 5 items. There are no subscales.

Populations.

The target population is patients with RA. The RADAI has been evaluated for use in patients with undifferentiated peripheral inflammatory arthritis (98). The RADAI-5 may be useful in other forms of inflammatory arthritis.

Practical Application

Method of administration.

Both the RADAI and RADAI-5 are self-administered questionnaires designed to be self-explanatory.

Administrative burden.

None.

Time to complete.

RADAI: patient, <5 minutes; provider, <30 seconds; laboratory, not applicable (N/A). RADAI-5: patient, <1 minute; provider, <30 seconds; laboratory, N/A (99).

Equipment needed.

A calculator or computer may be used to calculate the RADAI and RADAI-5.

Availability/cost.

The RADAI and RADAI-5 are available free of charge.

Scoring.

RADAI.

Items 1–3 are scored on numerical rating scales from 0 = “not at all” to 10 = “extremely/very severe.” Item 4 on morning stiffness is scored on an ordered category scale from 0 = “none,” 1 = “<30 minutes,” 2 = “30 minutes to 1 hour,” 3 = “1–2 hours,” 4 = “2–4 hours,” 5 = “>4 hours,” and 6 = “all day” and transformed to a 0–10 range. Item 5, the joint list, is calculated as the sum of 16 joints or joint groups scored from 0 indicating “no pain” to 3 indicating “severe pain,” with the 0–48 score transformed to the 0–10 range.

RADAI-5.

All items are scored on a 0–10 ordered category scale. For items 1 and 2, 0 = “completely inactive” and 10 = “extremely active.” For item 3, 0 = “no pain” and 10 = “unbearable pain.” For item 4, 0 = “very good” health status and 10 = “very bad” health status. For item 5, 0 = “no stiffness” and 10 = “stiffness the whole day.”

Both the RADAI and the RADAI-5 are calculated as the mean of the nonmissing items and range from 0–10.

Score interpretation.

Both RADAI and RADAI-5 scores of 0 mean no RA disease activity, with higher scores indicating higher levels of disease activity. Within individual patients, a change in RADAI >1.4 can be interpreted as clinically meaningful (38, 100).

Method of scoring.

The RADAI can be scored with a calculator or computer. The RADAI-5 can be scored by hand or with a calculator. There is no training required to interpret scores.

Norms available.

A RADAI score ≤2.0 has been identified as low or inactive disease (nonflare) (38); however, additional disease categories have not been established. Disease activity categories for the RADAI-5 have been proposed: 0.0–1.4 = a remission-like state, 1.6–3.0 = mild disease activity, 3.2–5.4 = moderate disease activity, and 5.6–10.0 = high disease activity (99).

Psychometric Information

Reliability.

Test–retest reliability.

Test–retest correlation for the RADAI performed within 1 week was high, with an interclass correlation coefficient of 0.92 (100). Test–retest reliability has not been proven for the RADAI-5.

Internal consistency.

For the RADAI, Cronbach's α of 0.87 and 0.91 have been reported (101, 102). For the RADAI-5, a Cronbach's α of 0.92 has been reported. Factor analysis by principal component analysis reveals both the RADAI and RADAI-5 to be 1-dimensional instruments with all items contributing significantly to the final scores (101, 103).

Validity.

Content.

The RADAI and RADAI-5 items score the most apparent symptoms of RA with the exception of joint scores. No physician-derived measures of disease activity or laboratory parameters are included in either measurement tool.

Criterion.

The RADAI is significantly correlated (P < 0.01) at low to moderate degrees with tender joint counts (TJCs; 0.44–0.55) and swollen joint counts (SJCs; 0.39–0.54), and variably correlated with acute-phase reactants (101, 102). The RADAI-5 is significantly correlated (P < 0.001) with TJCs (0.60) and SJCs (0.60); however, it was not correlated significantly with erythrocyte sedimentation rate or C-reactive protein (CRP) level (103). The κ for the relationship between the RADAI-5 and the original RADAI was excellent at 0.88, with fair agreement for the RADAI-5 as compared to the Disease Activity Score with 28-joint counts (DAS28; 0.29), DAS28-CRP (0.34), Simplified Disease Activity Index (SDAI; 0.39), and Clinical Disease Activity Index (CDAI; 0.37) (103).

Predictive.

The distribution of the RADAI is skewed toward normal, with up to 14% of patients scoring <1, preventing numerical improvement despite potential clinical improvement. Spearman's ρ demonstrated a linear relationship between change in RADAI-5 scores compared to change in DAS28 (0.59, P < 0.001) and change in CDAI (0.57, P < 0.001). Low to moderate κ correlations for the relationship between change in the RADAI-5 and change in the DAS28 (0.30) and change in the CDAI (0.54) have been demonstrated (103).

Convergent.

Spearman's rank correlations between the RADAI and RADAI-5 demonstrate near perfect correlation (0.995, P < 0.001). The RADAI is significantly correlated (P < 0.01) to a moderate degree with the Health Assessment Questionnaire (0.55–0.57) and DAS28 (0.53) (101, 102). The RADAI and RADAI-5 demonstrated correlations with the provider global assessment of disease activity (PrGA) of 0.54–0.59 and 0.60 (P < 0.01), respectively, compared to a correlation of 0.72 (P < 0.01) between PrGA and DAS28 (101–103). Statistically significant correlations between the RADAI-5 and DAS28, DAS28-CRP, SDAI, and CDAI were also found, with Spearman's ρ ranging from 0.64–0.74 (P < 0.001) (103).

Responsiveness to change.

The RADAI has been shown to detect flare of RA with equal sensitivity to the DAS28 (predictive ability 0.88) (38) and to discriminate European League Against Rheumatism good responders from moderate and nonresponders with predictive ability of 0.78 (100). For the RADAI-5, despite significant changes for individual patients, no significant changes were seen at the group level between 2 assessments done within 3 months (103).

Critical Appraisal of Overall Value to the Rheumatology Community

The RADAI and RADAI-5 are short, easy to understand, patient-administered tools that produce a single index of disease activity, designed for feasibility in a busy clinical setting. The RADAI was designed to provide a continuous score useful in clinical and epidemiologic research and is an aggregate of selected items from the 5 categories of the Rapid Assessment of Disease Activity in Rheumatology (RADAR) (101). The RADAR questionnaire is not discussed in this report as it does not provide a continuous score. The lack of physician, laboratory, and joint swelling assessments in the RADAI and lack of specificity to RA produced by the use of joint groups (i.e., fingers instead of metacarpophalangeal and proximal interphalangeal joints) detracts from the face validity of the measure. Additionally the RADAI has shown lower correlation with PrGA than DAS28 scores (102). The RADAI-5 improves upon the original RADAI in terms of simplicity of calculation, and the omission of joint counts makes assessment faster. Additionally, the disease activity categories of the RADAI-5 demonstrate a more normal distribution than DAS28 and CDAI scores (25). Despite lacking assessment of joint counts or physician-derived components, the RADAI-5 has been shown to correlate with these measures on an individual basis as well as with other composite measures that contain such items (103, 104). Relationships between both the RADAI and RADAI-5 and acute-phase reactant values have been weak or absent, diminishing the face validity of the composite measures (101–104). The RADAI lacks well-defined disease states, both the RADAI and RADAI-5 have not been validated outside of white cohorts (103), and the RADAI-5 has not been used in the research setting with sensitivity to change and relationships to disease outcomes, such as bony erosions or prediction of disability undocumented.

CHRONIC ARTHRITIS SYSTEMIC INDEX (CASI)

Description

Purpose.

The CASI combines single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. It is feasible to use for monitoring of RA disease activity in daily clinical practice.

Content.

The CASI includes the Ritchie Articular Index (RAI), patient assessment of pain visual analog scale (VAS), Health Assessment Questionnaire (HAQ), and erythrocyte sedimentation rate (ESR).

Developer/contact information.

G. F. Ferraccioli, MD, Division and Chair of Rheumatology, DPMSC, University of Udine, 33100, Udine, Italy. E-mail: gianfranco.ferraccioli@dpmsc.uniud.it (1993).

Versions.

None.

Number of items.

The CASI has 4 items. There are no subscales.

Populations.

The target population is patients with RA. It is not validated for other rheumatic disorders.

Practical Application

Method of administration.

Clinical assessment of joint counts and drawing a blood sample for ESR combined with self-administered patient assessments of disease activity and pain.

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

Patient, <3.5 minutes; provider, ∼3 minutes; laboratory, 1 hour waiting time for ESR.

Equipment needed.

A ruler may be required based on the VAS format chosen. A laboratory is needed for ESR determination. For calculation of the CASI, a calculator or computer may be used.

Availability/cost.

The CASI may be used free of charge; no specialized form is needed.

Scoring.

The RAI ranges from 0–78. The HAQ ranges from 0–3. The pain VAS ranges from 0–100. ESR generally ranges from 0–100. The range of the CASI is 0–74.

Score interpretation.

The level of disease activity can be interpreted as CASI remission ≤24.65; this corresponds to a Disease Activity Score (DAS) ≤3.32. Additional categories of disease activity have not been established (37).

Method of scoring.

The following formula is used: CASI = 13 × HAQ + 0.21 × ESR + 0.08 × pain VAS + 0.07692 × RAI. There is no training required to interpret the scores.

Norms available.

Other than remission, disease activity categories have not been established.

Psychometric Information

Reliability.

Test–retest.

The test–retest reliability of the CASI was tested after a period of 1 week with a test–retest correlation of 0.86 (105).

Internal consistency.

Cronbach's α of 0.81 has been reported (105).

Reproducibility.

The CASI was validated in a cross-sectional study inclusive of additional patients seen within the same clinic (37).

Validity.

Content.

The CASI includes variables from the American College of Rheumatology core set of measures used to assess outcomes in RA.

Concurrent.

One study over 6 months demonstrated a statistically significant decrease in CASI (39.9 to 24.7; P < 0.01), while at the same time demonstrating similar decreases in the HAQ (1.69 to 1.01; P < 0.05), ESR (55.2 to 34; P < 0.01), RAI (21.5 to 10.9; P < 0.01), and pain VAS (66.6 to 34.5; P < 0.01) (105). The CASI correlates with the DAS (R = 0.545, P < 0.01) (106) and with several indices of RA disease activity not currently in wide usage: the Thompson Articular Index (0.56, P < 0.01), the Lansbury Index (0.88, P < 0.01), the Stoke Index (0.44, P < 0.01), and the Mallya and Mace Index (0.53, P < 0.01) (105).

Construct.

The 4 factors composing the CASI contributed 74% of the common variance in a study of 124 patients: HAQ (39%), pain VAS (6%), ESR (8%), and RAI (21%). The CASI demonstrates a negative correlation with grip strength (0.44, P < 0.01) and positive correlations with number of swollen joints (0.39, P < 0.01), morning stiffness (0.42, P < 0.01), and the Zung Depression Inventory (0.52, P < 0.01) (105).

Responsiveness to change.

The CASI demonstrates similar sensitivity and specificity for remission as compared to the DAS (37). Over 6 months the CASI shows similar change over time as compared to ESR (106), C-reactive protein level, fibrinogen levels, morning stiffness, proximal interphalangeal joint synovitis score, pain levels tested by the Present Pain Index and the McGill Pain Questionnaire, the Performance and Activity Scale, and the Lansbury Index (P < 0.01), while changes in the Arthritis Impact Measurement Scales were nonsignificant in this population (105).

Critical Appraisal of Overall Value to the Rheumatology Community

The CASI was designed using factorial analysis of 29 available variables with the intent to design an RA measure of both disease activity and severity for use by practicing rheumatologists. Validation studies were mainly performed on measures no longer in general use. The inclusion of both patient- and provider-derived data as well as a laboratory marker of inflammation gives face validity to the CASI. The CASI correlates moderately with the DAS; however, the suggested cutoff for remission corresponds to a DAS much higher than the currently accepted remission level of 1.6. Patient global assessment of disease activity was not included in lieu of patient pain assessment as they were thought to measure the same thing, a claim supported by the high correlation between the 2 items (0.90, P < 0.01) (37). The calculation for the CASI requires use of a calculator or computer, which may make the measure difficult to use at point of care. The use of the RAI increases the time required to perform joint counts as compared to standard joint counts as grading of tenderness is required. Additionally, inclusion of ESR may make point of care use difficult in clinics that do not have laboratory values available at the time a patient is examined. The inclusion of the original HAQ increases the time required of the patient to complete the measure as compared to measures using shorter versions of the HAQ or alternative quality of life measures, and may not add sufficient additional information to justify the increase in time spent. Further studies are needed to define and validate categories of disease activity.

PATIENT-BASED DISEASE ACTIVITY SCORE WITH ESR (PDAS1) AND PATIENT-BASED DISEASE ACTIVITY SCORE WITHOUT ESR (PDAS2)

Description

Purpose.

The PDAS1 and PDAS2 combine single measures into an overall continuous measure of rheumatoid arthritis (RA) disease activity. The PDAS1 and PDAS2 are feasible to use for monitoring of RA disease activity in daily clinical practice.

Content.

The PDAS1 includes a patient-assessed 50–tender joint count (50TJC), patient global assessment of disease activity (PtGA), the modified Health Assessment Questionnaire (MHAQ), and erythrocyte sedimentation rate (ESR). The PDAS2 includes a patient-assessed 28–swollen joint count (28SJC), PtGA, the MHAQ, and an early morning stiffness (EMS) score.

Developer/contact information.

Ernest H. Choy, MS, ROCP, Sir Alfred Baring Garrod Clinical Trials Unit, Academic Department of Rheumatology, King's College London, Weston Education Centre, Cutcombe Road, London SE5 9RJ, UK. E-mail: ernest.choy@kcl.ac.uk (2008).

Versions.

The PDAS2 is similar in design to the PDAS1; however, the PDAS1 includes laboratory measurement of ESR, includes a patient-assessed 50TJC instead of a patient-assessed 28SJC, and excludes the EMS score.

Number of items.

The PDAS1 and PDAS2 each have 4 items. There are no subscales.

Populations.

The target population is patients with RA. It is not validated for other rheumatic disorders.

Practical Application

Method of administration.

Self-administered TJCs (PDAS1) or SJCs (PDAS2) using a mannequin displaying individual joints, patient self-assessments of disease activity, and physical function combined with laboratory evaluation of ESR (PDAS1) or self-administered EMS scale (PDAS2).

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

PDAS1: patient, 5–10 minutes (107); provider, <1 minute; laboratory: 1 hour waiting time for ESR. PDAS2: patient, 5–10 minutes (107); provider, <1 minute; laboratory, not applicable.

Equipment needed.

A ruler may be required based on the visual analog scale (VAS) format chosen. A laboratory is needed for ESR determination. For calculation of the PDAS1 and PDAS2, a calculator or computer is needed.

Availability/cost.

The PDAS1 and PDAS2 are available free of charge.

Scoring.

The 50TJC ranges from 0–50. The self-administered 28SJC ranges from 0–28. The PtGA ranges from 0–10. The MHAQ ranges from 0–3. The ESR generally ranges from 0–100. The EMS score ranges from 0–5. The range of the PDAS1 is 0–7.6. The range of the PDAS2 is 0–2.6.

Score interpretation.

The PDAS1 and PDAS2 produce continuous scores. Ranges for disease activity have not been established.

Method of scoring.

The following formulas are used: PDAS1 = 0.119 × (PtGA) + 0.842 × ln(ESR + 2) + 0.432 × ln(patient 50TJC + 2) + 0.271 × (MHAQ), and PDAS2 = 0.021 × (PtGA) + 0.483 × (MHAQ) + 0.033 × (self-administered 28SJC) + 0.002 × (EMS). There is no training required to interpret the scores.

Norms available.

Reference values have not been established.

Psychometric Information

Reliability.

During development of the tool, test–retest reliability was assessed for each item within 24 hours of initial assessment and was graded as excellent, with interclass correlation coefficients ranging from 0.76–0.88 for both the PDAS1 and PDAS2 (107).

Validity.

Content.

The PDAS1 includes 3 of the 7, and the PDAS2 includes 2 of the 7, American College of Rheumatology (ACR) core set measures used to assess outcomes in RA. While self-assessed joint counts are not included in the ACR core set of measures, self-assessed joint counts have been shown to variably correlate with joint counts performed by physicians (R = 0.41–0.78), but are not equivalent, with patients rating joint tenderness higher than physicians (60, 61, 108–110). Patient training has been shown to improve correlation with physician counts (R = 0.94) (110). In the development study, assessor 28TJCs demonstrated a correlation of 0.53 with the PDAS1 and a correlation of 0.76 with the Disease Activity Score with 28-joint counts (DAS28) (107).

Criterion.

The PDAS2 is less sensitive than both the PDAS1 and DAS28 for detecting low disease activity (107).

Construct.

The PDAS1 and PDAS2 showed convergent validity with moderate correlations with other measures of RA disease activity (PDAS1: DAS28 = 0.89, Clinical Disease Activity Index [CDAI] = 0.69, provider global assessment of disease activity [PrGA] VAS = 0.68, VAS = 0.72; PDAS2: DAS28 = 0.76, CDAI = 0.73, PrGA VAS = 0.67, VAS pain = 0.83). Both the PDAS1 and PDAS2 demonstrate divergent validity, with correlations between 0.23 and 0.37 for the PDAS1 and between 0.30 and 0.44 for the PDAS2 with the Nottingham Health Profile sleep, social, and emotion scales (107).

Responsiveness to change.

The PDAS1 and PDAS2 demonstrate similar sensitivity to change as compared to the DAS28, with a mean ± SD change over 6 months of 1.00 ± 1.30 for the PDAS1 and 0.73 ± 1.10 for the PDAS2 versus 1.20 ± 1.50 for the DAS28 (107).

Critical Appraisal of Overall Value to the Rheumatology Community

The PDAS1 and PDAS2 were designed to provide the information produced by the DAS28 in a simpler fashion in order to facilitate routine use in the clinic. As not all physicians perform detailed joint examinations when assessing RA disease activity, use of patient self-administered joint counts may be a good alternative to physician-administered joint counts (109). The format employed by the PDAS1 and PDAS2 may be superior to patient assessment of joints in list format (94, 107). The subjective nature of self-administered TJCs included in the PDAS1 may be easier for providers to accept than self-administered SJCs employed by some measures such as the PDAS2. Although simple patient training has been shown to improve quality of self-administered joint counts (110), patient training may impose additional burden on busy staff. Requirement of a 50-joint count in the PDAS1 is more extensive than most other currently used measures and may be cumbersome for patients, while not providing a significant amount of additional information. Additionally, lack of provider-derived information limits face validity of the measures. The PDAS1 requires laboratory measurement of ESR, which while adding face validity, may not be available at point of care in the clinical setting. The PDAS2 also uses an EMS scale, which is not used in many RA disease activity assessment tools at this time, despite evidence demonstrating that duration of morning stiffness has a moderate ability to differentiate active from inactive RA (111). Another drawback of both the PDAS1 and PDAS2 is that the complicated calculations, including transformation of variables, require use of a calculator. While a valid quality of life measure, and of the choices in HAQ versions, the MHAQ lacks a normal distribution, clusters at the lower end of the scale, and may fail to show numerical improvement in up to 25% of patients despite clinical improvement in function (82, 112–115), thus making the composite PDAS1 and PDAS2 less responsive at the lower end of the scale. Additionally, further studies are needed to define and validate categories of disease activity.

MEAN OVERALL INDEX FOR RHEUMATOID ARTHRITIS (MOI-RA)

Description

Purpose.

The MOI-RA combines single measures into an overall continuous measure of RA disease activity. It is feasible to use the MOI-RA for monitoring of RA disease activity in daily clinical practice.

Content.

The MOI-RA includes the mean of standardized values of tender and swollen joint counts (28-, 42-, or 66/68-joint counts), patient global assessment of disease activity (PtGA) on a 10-cm visual analog scale (VAS), patient assessment of pain on a 10-cm VAS, provider global assessment of disease activity (PrGA) on a 10-cm VAS, Health Assessment Questionnaire (HAQ), and erythrocyte sedimentation rate (ESR; range 0–100).

Developer/contact information.

Heidi Makinen, Jyvaskyla Central Hospital, Keskussairaalantie 19, 40620 Jyvaskyla, Finland (2008).

Versions.

The MOI-RA can be used with 28-, 42-, or 66/68-joint counts.

Number of items.

The MOI-RA has 7 items. There are no subscales.

Populations.

The target population is patients with RA. There are no other uses.

Practical Application

Method of administration.

Clinical assessment of joint counts and provider assessment of global disease activity, drawing a blood sample for ESR, and patient-administered HAQ, and pain and patient global assessment of disease activity VAS.

Administrative burden.

Training is necessary for reliable assessment of joint counts.

Time to complete.

Patient, 5–10 minutes; provider, 3–8 minutes; laboratory, 1 hour waiting time for ESR.

Equipment needed.

A ruler may be required based on the VAS format chosen. A laboratory is needed for ESR determination. For calculation of the MOI-RA, a calculator or computer is needed.

Availability/cost.

None.

Scoring.

All components are standardized to range from 0–100. VAS for PtGA, patient assessment of pain, and PrGA each range from 0–100. ESR values >100 are replaced by the value 100. HAQ values ranging from 0–3 are multiplied by 100 (28). The range of the MOI-RA is 0–100.

Score interpretation.

Higher values indicate poorer outcomes.

Method of scoring.

The mean of the standardized values is calculated producing a range for the MOI-RA between 0 and 100, with higher values indicating poorer outcomes. If values of 1–3 components are missing, standardized values are calculated from the available component values and the mean of the standardized values is recorded. The MOI-RA is most easily scored with use of a calculator. There is no training required to interpret scores.

Norms available.

Norms have not been documented for the MOI-RA.

Psychometric Information

Reliability.

Test–retest reliability has not been studied for the aggregate measure. The components of the MOI-RA are individually accepted as reliable in RA assessment.

Validity.

Content.

The MOI-RA is a composite measure inclusive of all 7 of the American College of Rheumatology (ACR) core set measures used to assess outcomes in RA chosen by expert consensus to represent valid dimensions of disease activity in RA.

Criterion.

The MOI-RA was compared to ACR response criteria with mean change in the MOI-RA from baseline to 6 months similar in patients grouped into those who did not meet ACR 20% response criteria (ACR20), those who met ACR20 but not ACR50 response, those with ACR50 but not ACR remission, and those who met ACR remission criteria. The MOI-RA was also compared to the Disease Activity Score with 28-joint counts (DAS28) with correlations between 0.84 and 0.90 over a period of 6 months (28).

Construct.

Changes in the MOI-RA were correlated with changes in the DAS28, which has been shown to correlate with disability over time (61) and contains the HAQ, which has itself demonstrated the ability to distinguish between placebo and treatment groups (116) with changes over time agreeing with and augmenting clinical and laboratory evidence of change (80, 117, 118).

Predictive.

No studies on predictive validity have been done. The MOI-RA has been shown to correlate with the DAS28 and HAQ, which have each been shown to predict RA-related morbidity (119, 120).

Responsiveness to change.

In the Finnish Rheumatoid Arthritis Combination Therapy (FIN-RACo) study, MOI-RA values decreased by ∼65% with a corresponding decrease in the mean DAS28 of 50%. Both the MOI-RA and DAS28 were able to significantly discriminate the 2 treatment arms of the FIN-RACo study (28).

Critical Appraisal of Overall Value to the Rheumatology Community

The MOI-RA was developed based on data for 169 RA patients with complete ACR core data; however, test–retest reliability of the composite measure has not been established. With inclusion of all 7 of the ACR core set measures, the MOI-RA has excellent face validity; however, it has not been widely used to this date, and as such the MOI-RA has not been further validated in additional RA cohorts. Disadvantages of the MOI-RA in clinical practice are the need for a blood sample, the time needed for providers to perform joint counts, and complicated mathematical calculation of the composite score. Additionally, cutoffs for categories of disease activity have not been established. ,

Table  . Summary Table for RA Outcome Measures*
ScalePurpose/contentMethod of administrationRespondent burdenAdministrative burdenScore interpretationReliability evidenceValidity evidenceAbility to detect changeStrengthsCautions
  • *

    RA = rheumatoid arthritis; DAS = Disease Activity Score; RAI = Ritchie Articular Index (53 joints in 26 graded units, graded for tenderness on pressure; maximum score 78); SJC = swollen joint count; ESR = erythrocyte sedimentation rate; PtGA = patient global assessment of disease activity; VAS = visual analog scale; DAS28 = Disease Activity Score with 28-joint counts; TJC = tender joint count; SDAI = Simplified Disease Activity Index; CRP = C-reactive protein; CDAI = Clinical Disease Activity Index; PrGA = provider global assessment of disease activity; N/A = not applicable; PAS = Patient Activity Score; HAQ = Health Assessment Questionnaire; PASII = Patient Activity Score-II; HAQII = Health Assessment Questionnaire-II; RAPID2 = Routine Assessment of Patient Index Data with 2 measures; RAPID3 = Routine Assessment of Patient Index Data with 3 measures; MDHAQ = Multidimensional Health Assessment Questionnaire; RAPID4PTJC = RAPID4-patient joint count; RAPID4MDJC = RAPID4-provider joint count; RAPID5 = Routine Assessment of Patient Index Data with 5 measures; RADAI = Rheumatoid Arthritis Disease Activity Index; RADAI-5 = Rheumatoid Arthritis Disease Activity Index-5; stiffness = early morning stiffness scale; CASI = Chronic Arthritis Systemic Index; MOI-RA = Mean Overall Index for Rheumatoid Arthritis; ACR = American College of Rheumatology; PDAS1 = Patient-Based Disease Activity Score with ESR; MHAQ = Modified Health Assessment Questionnaire; PDAS2 = Patient-Based Disease Activity Score without ESR.

  • Times were derived from multiple sources (authors of this article, published comments, and formal studies) in varying populations and should be considered approximate estimates. For the most part, they were not determined in head-to-head studies, precluding direct comparison between measures (7, 86, 88, 101, 121).

  • Psychometric properties were subjectively ranked using an ordered category scale (excellent, good, acceptable, poor, and unacceptable) based on currently published literature.

DAS
  • RA disease activity measurement

  • RAI: 0–78, SJC44: 0–44, ESR: 0–100, PtGA VAS: 0–100

Provider assessment, patient item, labPatient: ∼10 sec
  • Provider: 5–8 min

  • Lab: 1 hour waiting time for ESR

  • Remission: DAS <1.6

  • Low: 1.6 ≤ DAS < 2.4

  • Moderate: 2.4 ≤ DAS ≤ 3.7

  • High: DAS >3.7

ExcellentExcellentExcellent
  • Discriminates well between remission and active RA

  • Used extensively in clinical trials

  • Includes acute-phase reactant, patient- and provider-derived data

  • Requires provider joint counts, acute-phase reactant, calculator/computer for calculation

  • Heavily weighted for ESR

  • Not validated in non-RA disease states

DAS28
  • RA disease activity measurement

  • TJC28: 0–28, SJC28: 0–28, ESR: 0–100, PtGA VAS: 0–100

Provider assessment, patient item, labPatient: ∼10 sec
  • Provider: 3–5 min

  • Lab: 1 hour waiting time for ESR

  • Remission: DAS28 <2.6

  • Low: 2.6 ≤ DAS28 < 3.2

  • Moderate: 3.2 ≤ DAS28 ≤ 5.1

  • High: DAS28 >5.1

ExcellentExcellentExcellent
  • Discriminates well between remission and active RA

  • Used extensively in clinical trials

  • Includes acute-phase reactant, patient- and provider-derived data

  • Requires provider joint counts, acute-phase reactant, calculator/computer for calculation

  • Heavily weighted for ESR

  • Not validated in non-RA disease states

SDAI
  • RA disease activity measurement

  • TJC28: 0–28, SJC28: 0–28, PtGA VAS: 0–10, CRP: 0–10

Provider item, patient item, labPatient: ∼10 sec
  • Provider: ∼2 min

  • Lab: waiting time for CRP varies by lab

  • Remission: SDAI ≤3.3

  • Low: 3.3 < SDAI ≤ 11

  • Moderate: 11 < SDAI ≤ 26

  • High: SDAI >26

Good, test–retest reliability for composite has not been evaluatedExcellentExcellent
  • Simple mathematical scoring without transformation

  • Includes acute-phase reactant, patient- and provider-derived data

  • Requires provider joint counts, acute-phase reactant

  • Not validated in non-RA disease states

CDAI
  • RA disease activity measurement

  • TJC28: 0–28, SJC28: 0–28, PtGA VAS: 0–10, PrGA VAS: 0–10

Provider item, patient itemPatient: ∼10 sec
  • Provider: <2 min

  • Lab: N/A

  • Remission: CDAI ≤2.8

  • Low: 2.8 < CDAI ≤ 10

  • Moderate 10 < CDAI ≤ 22

  • High: CDAI >22

Good, test–retest reliability for composite has not been evaluatedExcellentExcellent
  • Simple mathematical scoring without transformation

  • Includes patient- and provider-derived data

  • All variables easily available at point of care

  • Requires provider joint counts

  • No acute-phase reactant included

  • Not validated in non-RA disease states

PAS
  • RA disease activity measurement

  • HAQ: 0–3; pain VAS: 0–10, PtGA VAS: 0–10

Patient questionnairePatient: <3.5 min
  • Provider: <1 min

  • Lab: N/A

  • Remission: ≤0.25

  • Minimal: ≤3.7

  • Moderate: <8.0

  • Severe: ≥8.0

Acceptable, test–retest reliability for composite has not been evaluatedAcceptableAcceptable for individual components, has not been evaluated for composite
  • Simple mathematical scoring without transformation

  • Includes patient-derived data only

  • No provider-derived data

  • Not validated in non-RA disease states

  • May be influenced by patient education level

PASII
  • RA disease activity measurement

  • HAQII: 0–3, pain VAS: 0–10, PtGA VAS: 0–10

Patient questionnairePatient: <1.5 min
  • Provider: <30 sec

  • Lab: N/A

  • Remission: ≤0.25

  • Minimal: ≤3.7

  • Moderate: <8.0

  • Severe: ≥8.0

Acceptable, test–retest reliability for composite has not been evaluatedAcceptableAcceptable for individual components, has not been evaluated for composite
  • Simple mathematical scoring without transformation

  • Includes patient-derived data only

  • No provider-derived data

  • Not validated in non- RA disease states

  • May be influenced by patient education level

RAPID2
  • Disease activity measurement in rheumatic disease

  • PtGA VAS: 0–10, PrGA VAS: 0–10

Provider item, patient itemPatient: 20 sec
  • Provider: <20 sec

  • Lab: N/A

Specific cutoffs not available, higher scores indicate greater disease activityAcceptable, test–retest reliability for composite has not been evaluatedPoorAcceptable for individual component, has not been evaluated for composite
  • Simple and easy to perform

  • Includes patient- and provider-derived data

  • No acute-phase reactant included

  • May be influenced by patient education level

  • Least information as compared to the other RAPID scores

RAPID3
  • Disease activity measurement in rheumatic disease

  • MDHAQ: 0–3, pain VAS: 0–10, PtGA VAS: 0–10

Patient questionnairePatient: ∼1.5 min
  • Provider: <30 sec

  • Lab: N/A

  • Remission: 0–1.0

  • Low: 1.1–2.0

  • Moderate: 2.1–4.0

  • High: 4.1–10

Acceptable, test–retest reliability for composite has not been evaluatedGoodAcceptable for individual component, has not been evaluated for composite
  • Has been used in all rheumatic diseases

  • Has demonstrated similar ability as compared to joint counts in differentiating active and inactive RA

  • Includes patient-derived data only

  • No provider-derived data

  • No acute-phase reactant included

  • May be influenced by patient education level

  • MDHAQ has a greater floor effect than original HAQ and less even spacing of questions than HAQII

  • Requires raw scores to be converted to 0–10 prior to addition of components and division by 3

RAPID4PTJC
  • Disease activity measurement in rheumatic disease

  • MDHAQ: 0–3, pain VAS: 0–10, PtGA VAS: 0–10, self-TJC: 0–48

Patient questionnaire, patient joint countPatient: 5–10 min
  • Provider: <1 min

  • Lab: N/A

  • Remission: 0–1.0

  • Low: 1.1–2.0

  • Moderate: 2.1–4.0

  • High: 4.1–10

Acceptable, test–retest reliability for composite has not been evaluatedGoodAcceptable for individual component, has not been evaluated for compositeIncludes patient-derived data only
  • No provider-derived data

  • No acute-phase reactant included

  • May be influenced by patient education level

  • MDHAQ has a greater floor effect than original HAQ and less even spacing of questions than HAQII

  • Requires raw scores to be converted to 0–10 prior to addition of components and division by 4

  • Lacks standardization

RAPID4MDJC
  • Disease activity measurement in rheumatic disease

  • MDHAQ: 0–3, pain VAS: 0–10, PtGA VAS: 0–10, TJC: 0–28, SJC: 0–26

Patient questionnaire, provider joint countPatient: <1.5 min
  • Provider: ∼2 min

  • Lab: N/A

  • Remission: 0–1.0

  • Low: 1.1–2.0

  • Moderate: 2.1–4.0

  • High: 4.1–10

Acceptable, test–retest reliability for composite has not been evaluatedGoodAcceptable for individual component, has not been evaluated for compositeIncludes patient- and provider-derived data
  • MDHAQ has a greater floor effect than original HAQ and less even spacing of questions than HAQII

  • No acute-phase reactant included

  • Requires raw scores to be converted to 0–10 prior to addition of components and division by 4

  • Lacks standardization

RAPID5
  • Disease activity measurement in rheumatic disease

  • MDHAQ: 0–3, pain VAS: 0–10, PtGA VAS: 0–10, PrGA VAS 0–10, TJC: 0–48

Patient questionnaire, provider item, patient or provider joint countPatient: 5–10 min if patient joint count and <5 min if provider joint count
  • Provider: <1 min if patient joint count and <2.5 min if provider joint count

  • Lab: N/A

  • Remission: 0–1.0

  • Low: 1.1–2.0

  • Moderate: 2.1–4.0

  • High: 4.1–10

Acceptable, test–retest reliability for composite has not been evaluatedGoodAcceptable for individual component, has not been evaluated for compositeMay include patient- and provider -derived data
  • No acute-phase reactant included

  • MDHAQ has a greater floor effect than original HAQ and less even spacing of questions than HAQII

  • Requires raw scores to be converted to 0–10 prior to addition of components and division by 5

  • Lacks standardization

RADAI
  • RA disease activity measurement, may be useful in other forms of inflammatory arthritis

  • Global activity over 6 months: 0–10, current activity: 0–10, pain: 0–10, health status: 0–6, self-TJC 0–48

Patient questionnairePatient: <5 min
  • Provider: <1 min

  • Lab: N/A

  • Proposed remission/low: 0–2

  • Other categories of disease activity not established

  • Higher values indicate poorer outcomes

ExcellentGoodGoodIncludes patient-derived data only
  • No acute-phase reactant included

  • No provider-derived data

  • Not validated in non-RA disease states

  • Validated only in whites

RADAI-5
  • RA disease activity measurement, may be useful in other forms of inflammatory arthritis

  • Global activity over 6 months: 0–10, current activity: 0–10, pain: 0–10, health status: 0–10, stiffness: 0–10

Patient questionnairePatient: <1 min
  • Provider: <30 sec

  • Lab: N/A

  • Remission: 0.0–1.4

  • Low: 1.6–3.0

  • Moderate: 3.2–5.4

  • High: 5.6–10.0

Good, test–retest reliability for composite has not been evaluatedGoodPoorIncludes patient-derived data only
  • No acute-phase reactant included

  • No provider-derived data

  • Not validated in non-RA disease states

  • Validated only in whites

CASI
  • RA disease activity measurement

  • HAQ: 0–3, pain VAS: 0–100, ESR: 0–100, RAI: 0–78

Patient questionnaire, provider assessment, labPatient: <3.5 min
  • Provider: ∼3 min

  • Lab: 1 hour waiting time for ESR

  • Remission: ≤24.65

  • Additional categories of disease activity not established

ExcellentGoodGoodInclusion patient- and provider-derived data and a laboratory marker of inflammation
  • Most validation studies performed on measures no longer in general use

  • Requires calculator/computer for calculation

MOI-RA
  • RA disease activity measurement

  • HAQ: 100 (0–3), PtGA VAS: 0–100, pain VAS 0–100, PrGA VAS: 0–100, ESR: 0–100

Provider item, patient item, labPatient: 5–10 min
  • Provider: 3–8 min

  • Lab: 1 hour waiting time for ESR

  • Categories of disease activity not established

  • Higher values indicate poorer outcomes

Good, test–retest reliability for composite has not been evaluatedGood, studies on predictive validity have not been doneExcellentIncludes all 7 ACR core set measures
  • Has not been validated outside inception cohort

  • Lacks standardization

PDAS1
  • RA disease activity measurement

  • MHAQ and self-TJC: 0–50, PtGA VAS: 0–10, ESR: 0–100

Patient questionnaire, labPatient: 5–10 min
  • Provider: <1 min

  • Lab: 1 hour waiting time for ESR

  • Categories of disease activity not established

  • Higher values indicate poorer outcomes

Acceptable, test–retest reliability for composite has not been evaluatedAcceptableAcceptable for individual componentsSelf-administered Includes acute-phase reactan
  • Requires calculator/computer for calculation

  • Patient-derived joint counts

  • Uses MHAQ, which has more pronounced floor effect than original HAQ, MDHAQ, or HAQII

  • Not validated for other rheumatic disorders

  • May be influenced by patient education level

PDAS2
  • RA disease activity measurement

  • MHAQ and self-SJC: 0–28, PtGA VAS: 0–10, stiffness: 0–5

Patient questionnairePatient: 5–10 min
  • Provider: <1 min

  • Lab: N/A

Higher values indicate poorer outcomesAcceptable, test–retest reliability for composite has not been evaluatedAcceptableAcceptable for individual components
  • Self-administered

  • Measures morning stiffness

  • Requires calculator/computer for calculation

  • Patient-derived joint counts

  • Uses MHAQ, which has more pronounced floor effect than original HAQ, MDHAQ, or HAQII

  • No acute-phase reactant included

  • Not validated for other rheumatic disorders

  • May be influenced by patient education level

PtGA VAS
  • Activity measurement for any disease state

  • 0–10 or 0–100 VAS

Patient itemPatient: ∼10 sec
  • Provider: ∼10 sec

  • Lab: N/A

  • Categories of disease activity not established

  • Higher values indicate poorer outcomes

  • Proposed low: ≤2.0

GoodAcceptableAcceptable
  • Self-administered

  • All aspects of disease important to patient considered

  • Lacks provider data

  • No acute-phase reactant included

  • May be influenced by patient education level

  • May include permanent damage

  • May include non-RA disease

PrGA VAS
  • Activity measurement for any disease state

  • 0–10 or 0–100 VAS

Provider itemPatient: N/A
  • Provider: ∼10 sec

  • Lab: N/A

  • Categories of disease activity not established

  • Higher values indicate poorer outcomes

  • Proposed low: ≤1.5 (0–10)

ExcellentGoodAcceptable
  • Provider administered

  • All aspects of disease important to provider considered

  • Lacks patient data

  • May include permanent damage

  • May include non-RA disease

Table  . Summary Table for Disease Activity Measures in Rheumatoid Arthritis*
 AssessmentRequire laboratory testPhysician joint countSelf-joint countPatient global VASProvider global VASHAQ versionMorning stiffnessPainDefined remission criteria
  • *

    VAS = visual analog scale; HAQ = Health Assessment Questionnaire; DAS = Disease Activity Score; N/A = not applicable; DAS28 = Disease Activity Score with 28-joint counts; SDAI = Simplified Disease Activity Index; CDAI = Clinical Disease Activity Index; PAS = Patient Activity Score; PASII = Patient Activity Score-II; HAQII = Health Assessment Questionnaire-II; RAPID2 = Routine Assessment of Patient Index Data with 2 measures; RAPID3 = Routine Assessment of Patient Index Data with 3 measures; MDHAQ = Multidimensional Health Assessment Questionnaire; RAPID4PTJC = RAPID4-patient joint count; RAPID4MDJC = RAPID4-provider joint count; RAPID5 = Routine Assessment of Patient Index Data with 5 measures; RADAI = Rheumatoid Arthritis Disease Activity Index; RADAI-5 = Rheumatoid Arthritis Disease Activity Index-5; CASI = Chronic Arthritis Systemic Index; MOI-RA = Mean Overall Index for Rheumatoid Arthritis; PDAS1 = Patient-Based Disease Activity Score with erythrocyte sedimentation rate (ESR); MHAQ = Modified Health Assessment Questionnaire; PDAS2 = Patient-Based Disease Activity Score without ESR; PtGA = patient global assessment of disease activity; PrGA = provider global assessment of disease activity.

  • While remission criteria for CASI, PtGA, and PrGA have been proposed, additional levels of disease activity are not defined.

DASProvider and patientYesYesNoYesNoN/ANoNoYes
DAS28Provider and patientYesYesNoYesNoN/ANoNoYes
SDAIProvider and patientYesYesNoYesYesN/ANoNoYes
CDAIProvider and patientNoYesNoYesYesN/ANoNoYes
PASPatientNoNoNoYesNoHAQNoYesYes
PASIIPatientNoNoNoYesNoHAQIINoYesYes
RAPID2Provider and patientNoNoNoYesYesN/ANoNoNo
RAPID3PatientNoNoNoYesNoMDHAQNoYesYes
RAPID4PTJCPatientNoNoYesYesNoMDHAQNoYesYes
RAPID4MDJCProvider and patientNoYesNoYesNoMDHAQNoYesYes
RAPID5Provider and patientNoNoYesYesYesMDHAQNoYesYes
RADAIPatientNoNoYesYesNoN/AYesYesNo
RADAI-5PatientNoNoNoYesNoN/AYesYesYes
CASIProvider and patientYesYesNoNoNoHAQNoYesYes
MOI-RAProvider and patientYesYesNoYesYesHAQNoYesNo
PDAS1PatientYesNoYesYesNoMHAQNoNoNo
PDAS2PatientNoNoYesYesNoMHAQYesNoNo
PtGAPatientNoNoNoYesNoN/ANoNoYes
PrGAProviderNoNoNoNoYesN/ANoNoYes

AUTHOR CONTRIBUTIONS

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.

ROLE OF THE STUDY SPONSOR

No author had any financial support or other benefits from commercial sources for the work reported on in the manuscript. Abbott Laboratories had no financial interest in this project and had no input in the design or content with all opinions and conclusions expressed herein those of the authors.

Ancillary