Abstract
Objective. To determine the prevalence of work disability in Spanish patients with ankylosing spondylitis (AS) and to identify factors related to it.
Methods. A cross-sectional study based on data from Regisponser (National Spanish Registry of Patients with Spondyloarthropathy). Demographic and disease-related variables were collected. AS patients were classified as work-disabled according to the Spanish Social Security System criteria. Variables that discriminated between AS patients with and those without work disability were identified using chi-square test or unpaired t test when appropriate. Multiple logistic regression was performed.
Results. In total 699 AS patients, age 48.7 ± SD 12.7 years and with disease duration 14.1 ± 10.1 years, were analyzed; 179 patients (25.6%) had permanent work disability. Several variables had significantly different values in patients with compared to those without work disability. In the regression model (pseudo R2 = 0.26, p < 0.0001), age (p = 0.001), sex (p = 0.04), disease duration (p = 0.006), total Bath AS Radiological Index (p = 0.007), Bath AS Functional Index (BASFI; p = 0.007), and chest expansion (p = 0.03) retained an independent association with work disability. When BASFI was excluded from the model the independent association with sex did not remain, and a significant association with finger to floor distance was found (p = 0.040).
Conclusion. The prevalence of permanent work disability in Spanish patients with AS is significant, and the main factors related to it are age, disease duration, structural damage, and physical functioning. Longitudinal studies are needed to confirm these results.
Ankylosing spondylitis (AS) is a chronic inflammatory disease that can cause several functional limitations, with significant impairment in patients’ ability to perform daily activities, including work tasks1. Indeed, work productivity is a relevant outcome with important consequences related to the patient’s health-related quality of life (HRQOL) and costs of the disease. Several studies have analyzed the influence of AS on patients’ ability to work2–11. Studies found a variable but significant prevalence of work disability, higher than expected in a matched general population3,9,11.
Regisponser is the National Spanish Registry of Patients with Spondyloarthropathies, including AS12. It contains data about more than 1000 patients, including disease-related variables and work outcomes, allowing analysis of work outcomes and related variables in Spanish patients with AS.
The aims of our study were to determine the prevalence of work disability in Spanish patients with AS and to identify the main factors related to it.
MATERIALS AND METHODS
This was a cross-sectional study based on data from the Regisponser. The registry was launched in 2004 by the Spondyloarthropathies Study Group of the Spanish Society of Rheumatology (GRESSER). Ten tertiary care centers accepted the invitation to participate in the registry. It contains data from patients who fulfilled the classification criteria of the European Spondylarthropathy Study Group (ESSG)13; all patients gave consent to be included in the registry. Collected data include demographic and disease-related and work-related variables.
Data available in April 2007 were used in this study. Patients with diagnosis of AS according to the modified New York criteria14 were studied. The primary outcome was the frequency of work disability, as defined by criteria of a damage evaluation unit of the Spanish Social Security System. Patients who had had a finding of “permanently work disabled” and received a disability pension were classified as work-disabled. Variables collected for purposes of this study included patient’s age, sex, disease duration, patient and physician global disease activity assessments measured using a visual analog scale (VAS) scored from 0 (very well) to 10 (very bad), patient pain assessment, and patient night pain assessment on a VAS from 0 (none) to 10 (maximum pain). Disease activity and physical functioning were assessed with the Spanish versions of the Bath AS Disease Activity Index (BASDAI)15 and the Bath AS Functional Index (BASFI)16. Metrology measurements included chest expansion, Schober test, occiput to wall distance, finger to floor distance, and lumbar side flexion. Structural damage was assessed by the Bath AS Radiology Index (BASRI)17, and spine-BASRI (s-BASRI) and total BASRI (t-BASRI) were calculated. HRQOL was assessed with the Spanish version of the AS Quality of Life scale (ASQOL; 0 = best HRQOL, 18 = worst HRQOL)18.
Statistical analysis
Descriptive statistics were performed. To identify variables that distinguished between AS patients with and without work disability, a comparison between work-disabled versus not work-disabled patients was performed using chi-square tests for categorical variables and unpaired t tests for continuous variables. A multiple logistic regression model was built up to identify factors associated with work disability. Associations between work disability and independent variables were first examined in univariate logistic regression analyses. Variables that had a significant association with work disability were included in the regression models as independent variables. In these models, the dependent variable was work disability. A p value < 0.05 was considered significant.
RESULTS
In total, 699 patients with AS, 535 men (77%), with age 48.7 ± SD 12.7 years and disease duration 14.1 ± 10.1 years, were included in the analysis. Of them, 179 (25.6%) had a permanent work disability. The main characteristics of the whole AS population and the subgroups with and without work disability are shown in Tables 1 and 2. As shown in these tables, age, sex, patient global assessment, patient pain and night pain assessments, BASDAI, BASFI, t-BASRI, s-BASRI, chest expansion and other metrology measurements, and HRQOL measured by the ASQOL discriminated between AS patients who were work-disabled and those who were not. The same variables had a significant association with work disability in the univariate logistic regression analyses (data not shown).
Characteristics of the total AS population and the subgroups with and without work disability (continuous variables).
Characteristics of the total AS population and the subgroups with and without work disability (categorical variables).
In the multiple logistic regression model (pseudo R2 = 0.26, p < 0.0001; Table 3) age, sex, disease duration, BASFI, t-BASRI, and chest expansion retained an independent association with work disability. When ASQOL was included in the model, it also retained an independent association with work disability (odds ratio 1.11, 95% CI 1.04–1.19, p = 0.001), but then the BASFI did not keep its association with work disability.
Multiple logistic regression model with work disability as independent variable.
We also constructed a multiple logistic regression model excluding the BASFI score. In this model (pseudo R2 = 0.25, p < 0.0001), age (OR 1.03, 95% CI 1.01–1.06, p < 0.0001), disease duration (OR 1.03, 95% CI 1.0–1.06, p = 0.006), t-BASRI (OR 1.4, 95% CI 1.2–1.7, p < 0.0001), chest expansion (OR 0.85, 95% CI 0.7–0.9, p = 0.019), and finger to floor distance (OR 1.01, 95% CI 1.00–1.03, p = 0.040) retained independent association with work disability.
DISCUSSION
In this cross-sectional study we assessed the frequency of work disability and the main variables related to it in a representative sample of Spanish patients with AS. We found that 25% of 699 patients with AS had permanent work disability. These findings are similar to previous reports, in which the prevalence of work disability ranges from 13% to 45%2–4,7,11, and are similar to the prevalence of work disability in a Spanish population with rheumatoid disease19. The variability among the studies can be related to several factors, including the characteristics of the AS population and the methodology used in each study (self-reported disability, having a disability pension, working status, etc.).
As the primary outcome measure of our study, we chose the judgment of “permanent work disabled” and “received a work disability pension” by the Spanish Social Security System. It is possible this underestimates the true prevalence of disability because those patients who were in sick-leave status were not considered as work-disabled. Moreover, patients who did not receive a disability pension in spite of being disabled (i.e., housewives and people without paid employment) were not considered as work-disabled in this study. In contrast, those patients who received a pension by reason of “partial disability” were taken into account. However, we think that our primary endpoint is a reliable measure that increases the robustness of the analysis in identifying factors related to work disability.
We were able to identify several variables related to work disability, mainly patient’s age, sex, disease duration, structural damage as measured by the BASRI, chest expansion, and physical functioning or HRQOL. Other studies have focused on identifying variables related to work disability in patients with AS2,3,6,7,10. Age, disease duration, physical functioning, and other work-related factors (i.e., physically demanding jobs) or coping styles have been related to work disability. Our study focused on demographic and disease-related variables without analyzing job-related factors as a cause of withdrawal from the labor force.
Age, sex, disease duration, and radiological damage were the AS-related variables found to be significantly associated with work disability in the multivariate analyses. Data concerning the age at which the work disability happened in each patient were not available. However, the association between age and work disability is well known and was consistent in our study.
Several metrology measurements (Schober test, chest expansion, occiput to wall distance, and finger to floor distance) were significantly related to work disability in the univariate analysis, but only chest expansion and finger to floor distance (in the model without BASFI) retained independent association with work disability in the multivariate analysis. Disease activity as measured by the BASDAI did not retain an association with work disability in the multivariate analysis.
When the ASQOL was included in the regression model, the association between work disability and BASFI did not remain, probably because physical functioning can be considered as a domain of HRQOL. Additionally, HRQOL should not be considered as a determinant of work disability because being work-disabled could probably contribute to the impairment of HRQOL. Thus, we chose the model without the ASQOL, in which the BASFI retained the association with work disability.
Because “physical functioning” and “work disability” can refer to similar domains we also performed the multivariate analysis without including the BASFI in the regression model. The results were similar, even though the independent association with patient’s sex did not remain, and a significant association was found between work disability and finger to floor distance. The t-BASRI, Schober test, finger to floor distance, and occiput to wall distance had significantly worse values in men than in women (data not shown). We think this could explain the association between work disability and patient’s sex in the model including the BASFI.
Taken together, our findings suggest that disease duration, physical functioning, and structural damage are the disease-related determinants of work disability in patients with AS. Patient’s sex is not likely to be an independent risk factor for work disability, and it is probable that work disability was more frequent in men than in women because men had more structural damage and worse metrology results than women.
From this perspective, the factors related to work disability should be modified in order to decrease the prevalence of work disability in patients with AS, but this must be confirmed in longitudinal studies. An early diagnosis of the disease and the ability of available therapies to prevent functional disability and structural damage should be key points in this strategy. Biologic agents have been proven to preserve the physical functioning of patients with AS, but there are many doubts about their ability to inhibit the radiographic progression. A recent study has shown a favorable effect of biologic therapy on ability to return to work in people with work disability due to AS8. In our study, only 52 patients (7.4%) were identified as receiving biologic therapy, and we were not able to find an association between treatment and work disability (data not shown).
The main limitation of our report is that we carried out a cross-sectional study, and so our findings should be considered with caution. Additionally, as discussed above, a unique outcome was considered (work disability according to the Spanish Social Security System criteria) and job-related and other factors such as coping styles were not analyzed. Specific information about the patient’s job type (physical vs non-physical) was not collected. Other variables such as time to disability, patient’s formal education, and previous treatments were not considered. These variables should be examined in future studies. The strength of the study is the large sample size, which yielded robust conclusions.
We observed a significant prevalence of work disability (25%) in Spanish patients with AS, and we were able to identify several variables related to it, mainly age, disease duration, structural damage measured by the BASRI index, and physical functioning or HRQOL. Longitudinal studies are needed to confirm these results.
Footnotes
-
Regisponser is supported by Schering-Plough, Madrid, Wyeth, Madrid, and Abbott Laboratories, Madrid, Spain.
- Accepted for publication June 29, 2009.
