Background/Purpose: Comorbidities can importantly influence the results of clinical studies on functional outcomes as they may act as confounders or effect modifiers. The generic self-administered comorbidity questionnaire (SCQ) is a frequently used instrument to assess common comorbidities which might impact functioning but has never been validated for use in ankylosing spondylitis (AS). Objective was to measure the agreement between SCQ-responses and medical records diagnosis, and to assess construct- and concurrent validity of the SCQ in patients with AS. 

Methods: The SCQ (range 0-45) asks about the presence, treatment and functional limitations of 12 common comorbidities and three additional non-specified medical problems. The SCQ, demographics and indices of disease activity (Bath AS Disease Activity Index [BASDAI]; AS Disease Activity Score- C-reactive protein); physical function (Bath AS Functional Index) and health-related quality of life (HR-QoL; Short form-36 [SF-36], Ankylosing Spondylitis Quality of Life [ASQoL], EuroQoL-VAS), were administered to 98 patients with AS who were followed in the Outcome in Ankylosing Spondylitis International Study (OASIS). The agreement between the SCQ-items and comorbidities retrieved from medical records was calculated by two independent extractors. Concurrent validity was assessed by the correlation with two other comorbidity indices: the Charlson index, a record-based comorbidity index which predicts mortality, and the Michaud/Wolfe index, which predicts functional outcomes. Construct validity was assessed by testing the hypothesis that a valid comorbidity index should correlate with age, function and overall HRQoL. An adapted version of the SCQ (adapted SCQ) was created and validated after removing items on rheumatic diseases (osteoarthritis, back pain, chronic rheumatic disease) because they were conceptually overlapping with the index disease. 

Results: The median SCQ-score was 5 (range 0-19) and the median adapted-SCQ-score was 2 (range 0-13). Frequently reported non-rheumatic comorbidities were hypertension (27.6%), inflammatory bowel disease (10.2%) and depression (9.2%). Agreement between self-report and medical records was moderate to perfect for all diseases included in the SCQ (kappa 0.47-1.00), except for stomach disease, depression, and osteoarthritis (kappa 0.14-0.15). The correlations of the SCQ with the Michaud/Wolfe index and the Charlson index were 0.39 and 0.24 respectively, and of the adapted-SCQ with both indices 0.53 and 0.36 respectively. The SCQ correlated weakly with age (r=0.24) and disease activity (BASDAI r=0.27), and moderately with function (r=0.43) and HRQoL (SF-36 physical r=-0.45; AsQoL r=0.43). The adapted SCQ correlated weakly with age (r=0.28), and moderately with function (r=0.41) and HRQoL (SF-36 physical r=-0.41, ASQoL r=0.32), but not with measures of disease activity.

Conclusion: The SCQ can be used to measure comorbidities which have impact on functional outcomes in AS, but the rheumatic items showed low agreement. Exclusion of these items improved construct and concurrent validity.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-the-self-administered-comorbidity-questionnaire-in-patients-with-ankylosing-spondylitis/