Clinical Rheumatology

, Volume 31, Issue 7, pp 1065–1071

RETRACTED ARTICLE: Quality of life of patients with rheumatoid arthritis in Argentina: reliability, validity, and sensitivity to change of a Spanish version of the Rheumatoid Arthritis Quality of Life questionnaire

Authors

  • Christian A. Waimann
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Fernando M. Dal Pra
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Maria F. Marengo
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Emilce E. Schneeberger
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Susana Gagliardi
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Jose A. Maldonado Cocco
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • Mónica Sanchez
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
  • A. Garone
    • Hospital General de Agudos “Dr. E. Tornú”
  • Rafael E. Chaparro del Moral
    • Hospital General de Agudos “Dr. E. Tornú”
  • Oscar L. Rillo
    • Hospital General de Agudos “Dr. E. Tornú”
  • Mariana Salcedo
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
    • Hospital General de Agudos “Dr. E. Tornú”
  • Javier E. Rosa
    • Hospital Italiano de Buenos Aires
  • F. Ceballos
    • Hospital Italiano de Buenos Aires
  • Enrique R. Soriano
    • Hospital Italiano de Buenos Aires
    • Instituto de Rehabilitación Psicofísica (IREP) de Buenos Aires
    • Chief Section of RheumatologyIREP
Original Article

DOI: 10.1007/s10067-012-1976-6

Cite this article as:
Waimann, C.A., Dal Pra, F.M., Marengo, M.F. et al. Clin Rheumatol (2012) 31: 1065. doi:10.1007/s10067-012-1976-6

Abstract

The Rheumatoid Arthritis Quality of Life (RAQoL) questionnaire is the first needs-based instrument specifically designed to measure quality of life (QoL) of patients with rheumatoid arthritis (RA). The aims of our study were to develop an Argentinean version of the RAQoL and to determine its reproducibility, validity, and sensitivity to change in patients with RA. Translation process was performed according to internationally accepted methodology. Internal consistency and test–retest reliability were calculated. Criterion and construct validity were assessed by comparing the RAQoL with parameters of disease activity, the Health Assessment Questionnaire (HAQ), and the Medical Outcomes Study 36-item health survey (SF-36) questionnaire. Sensitivity to change was measured at 6–12 months using standardized response mean (SRM). The minimal important change was defined as a change of 1 or 1.96 times the standard error of measurement. A total of 97 patients with RA were included. Cronbach’s α was 0.93, and test–retest reliability was 0.95. The RAQoL showed moderate to strong correlation with parameters of disease activity, the HAQ, and the SF-36. Functional status was the main determinant of patients’ level of QoL. The SRM of the RAQoL was 0.24. Agreement between 20 % improvement in RAQoL and ACR20 response was moderate. Minimal important change was 2.2 (1 SEM) or 4.3 (1.96 SEM). The Argentinean version of the RAQoL is the first Spanish translation of this questionnaire. Our findings show it to be valid, reliable, and sensitive to changes in RA clinical status.

Keywords

ArgentinaQuality of lifeRheumatoid arthritisSelf-questionnaireSpanish versionValidation

Introduction

Measurement of quality of life (QoL) is an increasingly important outcome in patients with rheumatoid arthritis (RA) [1]. Currently, the most widespread method for measuring QoL is self-reported questionnaires [25]. Their main advantages are low cost and easy implementation on any complexity level [6]. Classically, generic questionnaires have been used to measure QoL in patients with RA, including the Sickness Impact Profile [2], the Euro Quality of Life (EuroQol) [3], the Nottingham Health Profile(NHP) [4], the Medical Outcomes Study 36-item short form health survey (SF-36) [5], among others. However, these questionnaires are difficult to implement in daily practice because they are time-consuming, are difficult to understand, and have low sensitivity to changes in RA clinical status [7, 8].

The Rheumatoid Arthritis Quality of Life questionnaire (RAQoL) was specifically designed to measure QoL in patients with RA [9]. The content of the RAQoL was derived from interviews with patients with RA, and items were selected on the basis of unfulfilled needs as a result of having RA [10]. In this needs-based model, QoL is defined as “the extent to which individuals are able to meet their needs.” This model has been evaluated in different diseases and has been shown to have better psychometric properties than function-based instruments [11].

The RAQoL has been validated in diverse populations and languages and has been shown to have good reliability, validity, and sensitivity to change [10, 1223]. Its main advantages are its simplicity, rapid completion time, and greater sensitivity to change compared with generic questionnaires [8, 24]. The aim of our study was to develop an Argentinean version of the RAQoL and to determine its reproducibility, validity, and sensitivity to change in patients with RA.

Patients and methods

Design

We conducted a prospective cohort study including patients with RA to validate and evaluate sensitivity to change of an Argentinean version of the RAQoL.

Patients

Consecutive RA patients were recruited from a rheumatology hospital in Ciudad Autónoma de Buenos Aires, Argentina. All participants provided written informed consent. The Instituto de Rehabilitación Psicofísica ethics committee granted institutional approval. All participants were >18 years old, fulfilled the 1987 American College of Rheumatology (ACR) RA diagnostic criteria [25], and were proficient in the Spanish language.

Patients’ clinical and sociodemographic data were assessed at baseline. A second evaluation was performed at 6–12 months to assess sensitivity to change of the RAQoL. Collected data included age, sex, education level, income, employment status, health insurance status, self-reported pain and disease activity according to a visual analogue scale (VAS), physician-reported disease activity according to a VAS, 28-joint Disease Activity Score (DAS28) [26], Argentinean version of the Health Assessment Questionnaire (HAQ-A) [27], erythrocyte sedimentation rate, and rheumatoid factor serum levels.

Radiographic images of patients’ hands and feet were taken at baseline. The images were scored by one reader (blinded to patient clinical status and questionnaire scores) using the Sharp/van der Heijde method [28].

Instruments used in the study

  • The HAQ-A is a self-response questionnaire which is used to measure functional status. Subscale scores range from 0 to 3, with higher scores indicating worse functional status [27].

  • The Spanish version of the SF-36 is a generic health-related QoL instrument. It consists of 36 items aggregated into eight health domains: physical functioning, role limitations due to physical health, bodily pain, general health perceptions, vitality, social functioning, role limitations due to emotional problems, and mental health. These domains were scored on scales of 0–100, with higher scores indicating better health. In patients who had scores for all items, mental and physical health summary scores were calculated. These summaries were calculated using norm-based scoring methods, in which both summaries scales were standardized to a mean of 50 and a standard deviation of 10 in the general US population [5].

  • The RAQoL consists of 30 questions with yes/no response format. Each affirmative answer carries a score of one point. The total score is calculated as the sum of all the affirmative answers. Scores range from 0 to 30, with higher scores indicating poorer QoL [10].

Patients completed all questionnaires in the presence of their physician without assistance. Time to completion was assessed in only the first 24 patients.

Adaptation and validation methods

Adaptation and validation were carried out according to standard procedures [29]:
  • Translation and cross-cultural adaptation of the RAQoL were carried out according to internationally accepted methodology [30]. The original English version of the RAQoL was independently translated and adapted into local Spanish by three rheumatologists, and the final version was based on consensus. This final version was back-translated into English by a bilingual individual whose primary language was English. The back-translated version was compared with the original version, and discrepancies were highlighted and reviewed, arriving to the final version of The Argentinean-RAQoL questionnaire (“Appendix”).

  • Reliability assessed how reproducible are the results of the scale under different conditions. Two measures of reliability were used:
    • Internal consistency was assessed using Cronbach’s α coefficients, which indicated the extent to which the items of the scale are interrelated. A Cronbach’s α > 0.80 indicated an adequate interrelationship of items [31]. Additionally, an interpartial total correlation was calculated to evaluate the homogeneity of the scale. A correlation of 0.20–0.80 was considered good homogeneity.

    • Test–retest reliability was assessed in a group of patients who completed a second questionnaire 3–5 days after the first evaluation and whose clinical status remained stable according to physician reassessment. A correlation of ≥0.85 between the two administrations was accepted as good reproducibility, indicating that the instrument has a low level of random measurement error.

  • Construct validity was addressed using different comparators. According to previous publications, the RAQoL was expected to show a moderate to good positive correlation (correlation coefficients 0.40–0.80) with disease activity measures, self-reported scores for pain and activity, and functional status (HAQ-A score) [10, 13, 15, 17, 19, 20, 32, 33]. RAQoL was expected to be higher in patients with higher disease activity.

  • Criterion validity was measured using the SF-36 as a comparator instrument. Moderate to good negative correlations were expected between RAQoL scores and SF-36 scores, particularly in the domains measuring physical function, bodily pain, vitality, and role limitations due to physical health [12].

  • Sensitivity to change refers to “the ability of an instrument to measure change in a state…” [34] To assess sensitivity to change, a methodology similar to those reported in other studies was used [8, 24, 32].
    • Standardized mean response (SRM; mean change/standard deviation of change) was calculated for each outcome and classified according Cohen’s effect size: small (0.2), medium (0.5), and large (0.8) [35]. A large SRM indicated high sensibility to change. A subanalysis using patients with early arthritis (patients diagnosed with RA <2 years before study enrollment) was carried out to include those patients more prone to developing a change in RA clinical status.

    • At the second evaluation, patients were categorized as responders or non-responders according to ACR response criteria of 20 % (ACR20) and European League Against Rheumatism (EULAR) response criteria [36]. These criteria are considered the minimum clinically significant change in RA activity. Similarly, improvement of at least 20 % in their RAQoL score (RAQol-20 %) was calculated. Agreement between the RAQoL-20 %, ACR20, and EULAR criteria was measured using κ-statistics. Cohen’s κ coefficient was classified according to Landis and Koch’s criteria [37].

    • The minimal important change of RAQoL was determined using the standard error of measurement (SEM), which represents a true change above random measurement error. Two cutoff points to measure minimal important change have been described: the first using 1 SEM and a second, more conservative using 1.96 SEM. The SEM was estimated as SEM = (baseline standard deviation) × √(1 − R), where R is the reliability coefficient [38, 39].

Statistical analysis

Descriptive statistics were performed to calculate the means, standard deviations, medians, interquartile ranges, frequencies, and percentages. Non-normally distributed data were transformed into square roots or logarithms for multivariable analysis. RAQoL produces an ordinal level score, so nonparametric statistical tests were applied. Test–retest reliability, criterion validity, and construct validity were assessed using the Spearman rank correlation, and internal consistency was determined using Cronbach’s αcoefficient. Differences in RAQoL scores among groups were compared using the Mann–Whitney U test or Kruskal–Wallis one-way analysis of variance (where three or more independent groups were compared). Post hoc analysis was performed using the Mann–Whitney U test on each pair of groups and the P value was adjusted using the Bonferroni method.

Standard procedure was used to handle missing data from the HAQ-A, SF-36, and RAQoL [5, 10, 40]. In patients with missing items from the RAQoL, the total was multiple by 30 and then divided by the number of answered items. A total score was calculated only if no more than six items (20 % of the 30 items total) were missing.

Multivariate linear regression analysis using the RAQoL score as the dependent variable was performed to determine the impact of different variables on QoL. Variables with a significance level ≤0.15 in bivariate analysis were included as independent variables.

A two-sided P value of 0.05 was considered statistically significant. Data analyses were performed using STATA version 11.1 (Statacorp, College Station, TX, USA).

Results

Population characteristics

A total of 97 patients with RA were included in the study. Patients were primarily women (89 %), had a low level of literacy (59 % did not complete high school), and had a low income. Median disease duration was 72 months (range, 2–480 months). Thirty-seven patients (38 %) had early arthritis (disease duration <2 years). Table 1 shows the sociodemographic characteristics of the patients and their correlation with the RAQoL score.
Table 1

Patients’ baseline sociodemographic characteristics and correlation with RAQoL

 

n = 94

RAQoL Spearman rho’s coefficient

Age in years, mean ± SD

51.0 ± 12.7

−0.14

Women, n (%)

84 (89.4)

0.14

Diseases duration in months, mean ± SD

98.2 ± 106.4

−0.34b

Monthly income in US dollarsa, mean ± SD

445.2 ± 521.4

−0.06

Education level in years, mean ± SD

9.7 ± 4.0

−0.17

Patients with health insurance, n (%)

56 (59.6)

Unemployed, n (%)

39 (41.9)

SD standard deviation, n sample size

aData were missing for 34 patients

bCorrelation was significant at the <0.001 level (two-tailed)

Descriptive scores

The summary statistics for the RAQoL are described in electronic appendix (Table 1). The mean time taken to complete the questionnaire was 3 min. The percentage of missing answers was low (2.1 %), and only three questionnaires (3 %) were invalid due to the presence of ≥6 missing questions. These questionnaires were excluded from subsequent comparisons including the RAQoL.

Reliability

Cronbach’s α was 0.93, which indicated good internal consistency. All items presented a positive intertotal correlation (range, 0.38–0.66) with an average interitem correlation of 0.30 (range, 0.29–0.31), which indicated the absence of redundancy or poor interrelation between the items.

Test–retest reliability was assessed in 24 patients. The correlation was almost perfect (r = 0.95), which indicated good score reproducibility.

Criterion and construct validity

The RAQoL had a moderate to good correlation with physical and mental health summary scores of the SF-36, and the correlation was stronger in domains related to physical function (physical functioning and role limitations due to physical health). The RAQoL also showed moderate to strong correlation with disease activity and functional status measures (Table 2, electronic appendix). Figure 1 shows the ability of the RAQoL to distinguish between different levels of disease activity.
https://static-content.springer.com/image/art%3A10.1007%2Fs10067-012-1976-6/MediaObjects/10067_2012_1976_Fig1_HTML.gif
Fig. 1

Comparison of RAQoL median scores between different disease activity levels according to the DAS28 score. DAS28 28-joint Disease Activity Score, RAQoL Rheumatoid Arthritis Quality of Life, n sample size. Asterisk The RAQoL score of patients with high disease activity was significantly higher than the score of patients with moderate disease activity or disease remission

Sensitivity to change

A total of 79 patients (81 %) were reevaluated at 6–12 months (mean time = 8.9 ± 2.0 months). The SRMs of the measures at the second visit are shown in Fig. 2. All the measures indicated a small to medium effect size according to Cohen’s classification. The effect size of the RAQoL was small (SRM = 0.24), which was similar to that of the HAQ-A and lower than that of disease activity measures. In the early arthritis population, the SRM of the RAQoL was 0.54, which indicated a medium effect size. The level of agreement was fair between the RAQoL-20 % and EULAR criteria (κ = 0.34) and moderate between the RAQoL-20 % and ACR20 (κ = 0.45; Table 2). These results confirmed the ability of the RAQoL to measure change in RA clinical status.
https://static-content.springer.com/image/art%3A10.1007%2Fs10067-012-1976-6/MediaObjects/10067_2012_1976_Fig2_HTML.gif
Fig. 2

Standardized response mean (SRM) across the different outcomes. DAS28 28-joint index Disease Activity Score, RAQoL Rheumatoid Arthritis Quality of Life, TJC Tender Joint Count, SJC Swollen Joint Count

Table 2

Agreement between a 20 % of improvement in the RAQoL score and minimal clinically improvement according to ACR20 and EULAR response criteria

 

n

Respondersan (%)

Agreement with RAQoL-20 %

Observed agreement (%)

κ (95 % CI)

ACR20

70

21 (30.0)

73.53

0.45 (0.25–0.66)

EULAR

69

30 (43.5)

67.16

0.34 (0.11–0.56)

RAQoL-20 %

77

34 (44.2)

ACR20 American College of Rheumatology response criteria of 20 %, EULAR response reduction in DAS28 score ≥0.6, RAQoL-20 % criteria 20 % of improvement in RAQoL score, K Cohen’s kappa coefficient, 95 % CI 95 % confidence interval, n sample size

aPatients who achieved response according to the listed criteria

Minimal important change

The SEM of the Argentinean version of the RAQoL was 2.2. Thus, a change in the RAQoL score higher than 2.2 (1 SEM) or 4.3 (1.96 SEM) can be considered a minimal important change.

Determinants of QoL

Table 3 shows the results of multivariate linear regression analyses that were performed to determine the independent influence of clinical and demographic characteristics on RAQoL score. HAQ-A was the main variable associated with a reduction in QoL (β = 0.65; R2 = 0.61). A second model without HAQ-A showed a statistically significant association between RAQoL and DAS28 and self-reported levels of pain (β = 0.26; β = 0.36; R2 = 0.40).
Table 3

Multiple linear regression using the Argentinean version of the RAQoL score as the dependent variable

Predictors

Model 1. Including HAQ-A

Model 2. Without HAQ-A

β

p

β

p

Patient VAS of pain

0.08

>0.20

0.36

0.001

DAS28

0.01

>0.20

0.26

0.02

Years of education

−0.10

0.19

−0.14

0.12

Age

−0.10

0.19

−0.10

>0.20

Female gender

0.07

>0.20

0.08

>0.20

Diseases duration

−0.09

>0.20

−0.05

>0.20

HAQ-A

0.65

<0.001

R2

0.61

0.40

HAQ-A Health Assessment Questionnaire Argentinean version, VAS visual analogue scale, DAS28 28-joint Disease Activity Score

Discussion

In this study, we found that the Argentinean version of the RAQoL is a valid and reliable measure of QoL in patients with RA. Its main advantages are its simplicity, the easy calculation of its scores, and rapid completion time which allow its use in daily clinical practice. Almost all patients completed the questionnaire (only three invalid questionnaires), despite the low literacy level of the study population. The RAQoL was able to distinguish between different levels of disease activity and changes in RA clinical status. This finding suggests that the RAQoL is a good instrument for comparing the effectiveness of drugs, treatment strategies, and physical interventions. The sensitivity to change of the RAQoL was similar to that reported in other populations (range, 0.20–0.69) and was shown to be better than that of generic self-questionnaires of QoL (SF-36, EuroQol, and NHP) [8, 24, 32].

The minimal important change was 2.2 points with a 1-SEM cutoff or 4.3 points with a 1.96-SEM cutoff. This finding is consistent with values reported in other validations [13].

The HAQ-A was the most important determinant of QoL. However, the model including HAQ-A scores could explain only 61 % of the variance in the RAQoL score, which suggests that physical function is not the only determinant of change in the score. This strong relationship between HAQ-A and RAQoL was also found in others validations [13, 15, 17, 19, 24, 32, 33].

The adaptation and validation of this version of the RAQoL was carried out according to standard procedures [29]. However, a limitation of our study was that a Rasch analysis was not performed [41]. This analysis would have allowed a more comprehensive evaluation of the unidimensionality and interval levels of the scales [15].

Recently, Pacheco-Tena et al. developed and validated a Mexican version of the RA-QoL. That version showed good reliability and reproducibility. The questionnaire was able to discriminate between different levels of patients self-reported diseases activity and general health. However, sensibility to change and correlation with RA composite scores (e.g., DAS28, ACR) was not evaluated [42].

In conclusion, the RAQoL is the first needs-based QoL instrument specifically designed for patients with RA [10]. This version showed to be valid, reliable, and sensitive to changes in RA clinical status.

Key points

QoL is an important outcome in patients with RA. The RAQoL is the first needs-based instrument specifically designed to measure QoL in patients with RA. The Argentinean version of the RAQoL is valid, reliable, and sensitive to changes in RA clinical status.

Acknowledgments

This study was supported by an unrestricted grant from Wyeth-Pfizer, Argentina.

Disclosure

Christian A. Waimann has received honoraria from Wyeth-Pfizer, Argentina for this work. All other authors declare no conflicts of interest.

Supplementary material

10067_2012_1976_MOESM1_ESM.doc (52 kb)
ESM 1(DOC 52 kb)

Copyright information

© Clinical Rheumatology 2012