Evaluation of the psychometric properties of the St George's Respiratory Questionnaire in patients with severe asthma

Abstract

PurposeLimited data exist on the quantitative validity of the St George's Respiratory Questionnaire (SGRQ) in asthma populations. This study evaluated the psychometric properties of the SGRQ in patients with severe asthma. MethodsThis was a post-hoc analysis of pooled data from MENSA (N = 576; NCT01691508) and SIRIUS (N = 135; NCT01691521), two randomized, placebo controlled trials of mepolizumab in patients with severe asthma. Patients completed the SGRQ at Baseline and Exit (MENSA Week 32; SIRIUS Week 24). Distributional characteristics, internal consistency reliability, test-retest reliability, convergent and discriminant validity, known-groups validity and responsiveness were assessed. ResultsInternal consistency reliability was acceptable for the total and domain scores at Baseline and Exit (Cronbach's α was 0.92 and 0.94 at Baseline and Exit, respectively, for the total score). Test-retest reliability was demonstrated (intraclass correlation coefficients >0.7) for total score and the Activity and Impacts domains. Convergent and discriminant validity were demonstrated with measures associated or not associated with respiratory-related health status. Known groups validity based on baseline FEV1% predicted, Asthma Control Questionnaire (ACQ)-5 score, exacerbations and eosinophil counts was demonstrated for the SGRQ total and domain scores. Responses to therapy based on clinician-rated response, patient-rated response, ACQ-5 change score and exacerbations generally correlated with improvements in SGRQ scores. ConclusionsThis analysis demonstrated that the SGRQ has acceptable psychometric properties in patients with severe asthma, exceeding the thresholds for adequate reliability, validity and responsiveness. The SGRQ appears to be a good instrument for identifying response to therapy in patients with severe asthma.