Patient self-report RADAI (Rheumatoid Arthritis Disease Activity Index) joint counts on an MDHAQ (Multidimensional Health Assessment Questionnaire) in usual care of consecutive patients with rheumatic diseases other than rheumatoid arthritis
New York University School of Medicine and NYU Hospital for Joint Diseases, New York
To analyze a patient self-report joint count from the Rheumatoid Arthritis Disease Activity Index (RADAI) on a Multidimensional Health Assessment Questionnaire (MDHAQ) in a cohort of consecutive patients seen in usual rheumatology care with rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), osteoarthritis (OA), psoriatic arthritis (PsA), and gout.
Each patient completed an MDHAQ, which included a RADAI, at each visit in one usual care setting. In order to include a physician measure, a random visit at which there was a recorded physician global estimate was selected for each of 465 patients (174 patients with RA, 75 with SLE, 113 with OA, 53 with PsA, and 50 patients with gout). The RADAI was analyzed for total scores (range 0–48), number of involved joint groups (range 0–16), and each specific joint group, and then compared in the 5 diagnostic groups to one another and to other MDHAQ measures and the Routine Assessment of Patient Index Data 3 (RAPID3).
In patients with RA, SLE, OA, PsA, and gout, mean RADAI scores (range 0–48) were 12.4, 6.5, 10.1, 6.7, and 2.7, respectively. The mean numbers of involved joint groups (range 0–16) were 6.9, 3.8, 4.8, 4.5, and 1.7, respectively, and the median numbers were 6, 2, 4, 4, and 1, respectively. RADAI scores were correlated significantly with the physician global estimate, except in SLE, and at higher levels with the MDHAQ and RAPID3 scores in all diagnostic groups.
The RADAI self-report joint counts can be used to record self-report involvement of specific joints and joint groups in patients with SLE, OA, PsA, and gout, with minimal effort on the part of the rheumatologist.
A self-administered joint count was reported initially as a Rapid Assessment of Disease Activity in Rheumatology questionnaire by Mason et al (1) and extended as a Rheumatoid Arthritis Disease Activity Index (RADAI) in 1995 by Stucki et al (2). RADAI scores are correlated significantly with tender and swollen joint counts in rheumatoid arthritis (RA) patients (3–7). Although developed initially for RA, the authors suggested that the RADAI could be informative in patients with other rheumatic diseases. However, self-report joint counts have not been studied systematically in rheumatic diseases other than RA (3–7) except for one report regarding ankylosing spondylitis (8).
A RADAI self-report joint count was introduced into a Multidimensional Health Assessment Questionnaire (MDHAQ) in 2005 (9). Successful collection of an MDHAQ in patients with RA at each visit generally involves distribution to every patient, with any diagnosis, at every visit. This practice has been in place at the Seligman Center for Advanced Therapeutics of the NYU Hospital for Joint Diseases since July 2005. Completion of an MDHAQ, including RADAI scores, appeared of value in almost all patients. Therefore, a study was conducted to analyze how informative RADAI scores might be in patients with systemic lupus erythematosus (SLE), osteoarthritis (OA), psoriatic arthritis (PsA), or gout, as well as RA. In this report, we analyze RADAI self-report joint counts in consecutive patients with 5 diagnoses to study whether they add incremental information to the other scales on the MDHAQ.
Significance & Innovations
The Rheumatoid Arthritis Disease Activity Index (RADAI), a patient self-report of painful joint groups, documents involvement of specific joint groups in patients with systemic lupus erythematosus (SLE), gout, osteoarthritis (OA), and psoriatic arthritis (PsA), as well as in rheumatoid arthritis (RA), adding incremental information to other Multidimensional Health Assessment Questionnaire (MDHAQ) scales in many patients.
RADAI scores indicate self-report of involvement of >4 joint groups in 64% of patients with RA, 33% with SLE, 58% with OA, 57% with PsA, and 18% with gout, suggesting that further investigation of the basis for these findings would be of value.
RADAI scores are correlated significantly with physician global estimate in patients with RA, PsA, OA, and gout, but not SLE, and at a higher level with other MDHAQ scores for physical function, pain, patient global estimate, and the Routine Assessment of Patient Index Data 3 in patients with all 5 diagnoses.
Patients and methods
All patients seen at the Seligman Center for Advanced Therapeutics of the NYU Hospital for Joint Diseases complete an MDHAQ (with a RADAI self-report joint count) at every visit while waiting to see the rheumatologist in the infrastructure of care. All MDHAQ scores, as well as laboratory tests and medications, are recorded in a database for usual care. The database included 6,268 visits of 1,515 patients with RA, SLE, OA, PsA, or gout between July 2005 and April 2011. In order to include a quantitative clinical measure by a physician (as formal joint counts are not performed in SLE, OA, and gout, and only inconsistently in PsA and RA), only visits that included a physician global estimate of status (DOCGL) were selected for the study. A DOCGL was available in 465 of the 1,515 patients, including 174 RA patients, 75 SLE patients, 113 OA patients, 53 PsA patients, and 50 gout patients. A random visit of each of these 465 patients was selected for this study. No selection criteria determined which patients had a DOCGL recorded. The diagnosis was established by the treating physician.
Patient self-report MDHAQ.
The MDHAQ (9, 10) includes 10 queries concerning activities, each of which is scored 0–3. Scoring options are 0 (without any difficulty), 1 (with some difficulty), 2 (with much difficulty), and 3 (unable to do) for a total physical function (FN) score of 0–30. The raw 0–30 score is converted to 0–10 using a template on the MDHAQ. Visual analog scales (range 0–10) for pain and patient global estimate of status (PATGL) are positioned above and below a self-report joint count based on the RADAI (2). The Routine Assessment of Patient Index Data 3 (RAPID3) (11) is a composite of 3 measures of FN, pain, and PATGL (total score 0–30) (11). Demographic data on the MDHAQ include sex, date of birth, ethnicity, and years of formal education (9, 10).
RADAI self-report joint count.
The RADAI self-report joint count queries patients to score pain in 16 specific joint groups, 8 each on the right and left sides: fingers, wrists, elbows, shoulders, hips, knees, ankles, and toes. Scoring options are 0 (no pain), 1 (mild pain), 2 (moderate pain), or 3 (severe pain). The total score range is 0–48 (2). The RADAI joint count on the MDHAQ also includes the back and neck, for which responses are helpful clinically, but are not included in total scores. The full RADAI includes visual analog scales for pain, global estimate, and morning stiffness (2), but similar scales are found on the MDHAQ (9), and RADAI scales are not repeated to avoid redundancy.
In this report, in addition to total RADAI joint count scores based on a 0–3 rating of 16 joint groups for a total of 0–48, totals are also calculated for the number of joint groups involved (scored as 1) or not involved (scored as 0) for a total range of 0–16. This calculation was performed to simulate a standard physician formal 28-joint count, in which only normal/abnormal, rather than a graded score, is recorded (12).
Physician and laboratory measures.
As noted, a DOCGL was required for inclusion in the study and was available for all 465 study patients. Erythrocyte sedimentation rate (ESR) was available in 193 of these patients, and C-reactive protein (CRP) level was available in 151 patients. No selection criteria were in place for patients in whom these measures were available or not.
The MDHAQ is incorporated into the medical record for clinical care. The data are entered into a Microsoft Access database, which includes demographic information, MDHAQ scores, medication, and laboratory data for all patients seen. Approval by the Institutional Review Board of New York University School of Medicine was obtained for a retrospective chart review of the data, which had been collected prospectively in usual patient care.
For this study, a database was extracted from the Microsoft Access database to include a random visit of each of the 465 patients for whom a DOCGL was recorded, as noted above. The database was transferred to STATA software, version 12.0 for Windows, for all analyses.
Means, medians, and frequencies of RADAI scores were computed for patients with each diagnosis as total scores (range 0–48) and as the number of involved joint groups (range 0–16), in order to simulate a physician joint count (in which only normal/abnormal rather than a graded score is recorded) (12). The proportion of patients scoring each specific joint group as painful was also computed for each diagnostic group. As the RADAI has been validated in RA, regressions were computed to compare the mean number of painful joint groups, adjusted for age and sex, in each diagnostic group to RA.
Construct validity of RADAI scores was analyzed through correlations with DOCGL and other patient self-report MDHAQ scores for FN, pain, PATGL, and RAPID3 as well as ESR and CRP level. Most variables were not normally distributed or had ordinal characteristics, so Spearman's correlation coefficients were used. Interpretation of statistical significance was adjusted for multiple comparisons (13).
The 465 patients were 68% women, including 89% of SLE patients, 73% of OA patients, 35% of PsA patients, 20% of gout patients, and 80% of RA patients (Table 1). Fifty-nine percent of the patients were white, including slightly fewer than 50% of patients with RA or SLE. The mean ± SD level of formal education was 15.2 ± 3.5 years. The mean disease duration was 6.7 years (Table 1).
Table 1. Demographic and clinical characteristics of patients according to diagnosis*
All patients (n = 465)
Patient groups by diagnosis
RA (n = 174)
SLE (n = 75)
OA (n = 113)
PsA (n = 53)
Gout (n = 50)
RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; OA = osteoarthritis; PsA = psoriatic arthritis; PATGL = patient global estimate of status; RAPID3 = Routine Assessment of Patient Index Data 3; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; DOCGL = physician global estimate of status.
P < 0.0001 for each diagnosis compared to RA (analysis of variance for continuous variables and chi-square test for categorical variables).
Mean MDHAQ FN scores (range 0–10) in all patients were 2.0, including 2.6 in RA patients, and ranged from 1.1 (SLE) to 2.3 (OA) in other patients. Mean pain scores (range 0–10) in all patients were 4.3, including 4.9 in RA patients, and ranged from 2.8 (SLE) to 5.3 (OA) in other patients. Mean PATGL scores (range 0–10) were 4.0, including 4.6 in RA patients, and ranged from 2.5 (gout) to 4.6 (OA) in other patients. Mean RAPID3 scores (range 0–30) were 10.2, including 12.2 in RA patients, and ranged from 6.8 (gout) to 12.1 (OA) in other patients. Mean DOCGL scores (range 0–10) were 2.8, including 2.9 in RA patients, and ranged from 1.8 (gout) to 3.0 (PsA) in other patients.
RADAI self-report painful joint score.
In the RA, SLE, OA, PsA, and gout patient groups, mean RADAI total scores (range 0–48) were 12.4, 6.5, 10.1, 6.7, and 2.7, respectively, and the median total scores were 9 (range 2–20), 2 (range 0–5), 7 (range 4–13), 5 (range 2–9), and 1 (0–4), respectively (Table 2). Mean numbers of involved joint groups (range 0–16) were 6.9, 3.8, 5.8, 4.5, and 1.7, respectively, and median numbers were 6 (range 2–12), 2 (range 0–5), 4 (range 3–8), 4 (range 1–7), and 1 (range 0–2), respectively, in patients in the 5 diagnostic groups. (Table 2). An abnormal RADAI score >0 was reported by 99% of patients with OA, 87% of patients with RA, 83% with PsA, 60% with gout, and 59% with SLE (Table 2). Involvement of ≥4 joint groups was self-reported by 52% of all patients, including 64% with RA, 33% with SLE, 58% with OA, 57% with PsA, and 18% with gout (Table 2).
Table 2. RADAI self-report joint count on the MDHAQ in patients with 5 rheumatic diseases*
RADAI joint count scores
All patients (n = 465)
Patient groups by diagnosis
RA (n = 174)
SLE (n = 75)
OA (n = 113)
PsA (n = 53)
Gout (n = 50)
Values are the number (percentage) of patients unless indicated otherwise. RADAI = Rheumatoid Arthritis Disease Activity Index; MDHAQ = Multidimensional Health Assessment Questionnaire; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; OA = osteoarthritis; PsA = psoriatic arthritis; IQR = interquartile range.
P < 0.0001 for each diagnosis compared to RA.
Joint groups with 2 highest proportions of involvement in each patient group.
The mean number of involved joint groups, adjusted for age and sex, differed significantly in RA versus SLE and gout (Figure 1) (P < 0.001, adjusted for multiple comparisons). No significant differences were seen between RA versus PsA or OA (P > 0.05, adjusted for multiple comparisons). The 95% confidence intervals in patients with RA did not overlap with those of patients with SLE and gout (Figure 1).
The joint groups self-reported as the 2 most affected in all patients were knees (47%) and fingers (43%) (Table 2). Fingers were among the 2 most affected joint groups in patients with all diagnoses except gout. Knees were among the 2 most affected joint groups in all diagnoses except SLE. Among other joint groups, wrists were self-reported as most affected in RA, shoulders in SLE, and toes in gout (Table 2).
RADAI scores were correlated significantly with DOCGL in RA, PsA, OA, and gout (data not shown; ρ = 0.25–0.54, P < 0.05, adjusted for multiple comparisons), but not in SLE, perhaps reflecting that articular findings are less prominent in SLE than in the other diagnoses. Correlations of RADAI scores with MDHAQ FN, pain, PATGL, and RAPID3 were significant in all diagnoses (data not shown; ρ = 0.49–0.77, P < 0.01 for all, adjusted for multiple comparisons). Correlations between RADAI joint counts and laboratory tests were not statistically significant, even without adjustment for multiple comparisons (data not shown).
To our knowledge, this is the first report concerning self-report joint counts in patients with rheumatic diseases, other than in RA (3–7) and a single report concerning ankylosing spondylitis (8). The RADAI indicates self-reported involvement of specific joint groups in any patient with a rheumatic disease. The 95% confidence limits of adjusted mean number of involved joint groups for RA did not overlap with SLE or gout. Therefore, the RADAI provides incremental information to other MDHAQ scales for FN, pain, PATGL, and RAPID3.
RADAI data indicated self-report of ≥4 involved joint groups in 58% of patients with OA and 57% with PsA, compared to 64% with RA. Median RADAI scores were 7 in OA and 5 in PsA versus 9 in RA, and median numbers of involved joint groups were 4 in OA and PsA versus 6 in RA. The unexpectedly high self-reported joint involvement in patients with OA may reflect better results of treatment of RA versus OA. Nonetheless, even 33% of patients with SLE and 18% with gout self-reported ≥4 involved joint groups.
The only quantitative measure by the physician applicable to patients with diagnoses other than RA and PsA was a physician global estimate, as formal joint counts are not performed in patients with SLE, OA, or gout. Significant correlations were seen between RADAI self-report joint scores and physician global estimates, except in SLE patients. This finding may reflect in part that joint involvement plays a lesser role in overall patient status in SLE than in the other diseases studied.
This study includes several limitations. First, as noted, joint counts or formal examinations by a physician were not performed for comparison with RADAI self-report joint counts; potential disagreement may exist between patient and clinician estimates of joint involvement, which was not analyzed in this study. Second, this initial report cannot provide information concerning whether self-report of joint involvement may reflect inflammation (which is likely in most patients with RA, PsA, SLE, and gout), damage in most patients with OA and some joints in patients with other diagnoses, and possible fibromyalgia in any diagnosis, which indicates a need for further research, including evaluations by health professionals. Third, longitudinal data are not presented to recognize whether RADAI scores might document improvement or worsening of patient status over time.
Nonetheless, a RADAI self-report joint count in routine patient assessment appears preferable to the current absence of any quantitative articular data in the medical records of most patients with rheumatic diseases other than RA, and even of most RA patients (14). At this time, the only quantitative data in most medical records of patients seen by rheumatologists often are laboratory tests, which usually are not available at the time of the visit. Patient self-evaluation is far less time-consuming for the physician than a formal joint count. Patient self-report joint counts can overcome, in part, variations seen between different examinations done by a single individual and especially among different assessors (15).
Joint counts and evaluations performed by health professionals cannot be replaced by RADAI self-report joint scores. Formal comparisons of RADAI self-report scores to formal joint counts and joint examinations by physicians are needed to clarify the basis for self-report of joint involvement in different diseases. RADAI self-report joint counts can provide quantitative data at all patient visits concerning specific articular involvement of joints and joint groups in all patients with all rheumatic diseases.
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 submitted for publication. Dr. Pincus had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Castrejón, Yazici, Pincus.
Acquisition of data. Yazici.
Analysis and interpretation of data. Castrejón, Yazici, Pincus.
ROLE OF THE STUDY SPONSOR
Bristol-Myers Squibb and UCB Pharma had no role in the study design, data collection, data analysis, data interpretation, writing of this report, decision to publish, or approval of the version to be published.