We found that neither SGRQ total nor dimensional score were affected by sex. There was no statistically significant correlation between age and SGRQ dimensions and total scores. SGRQ total and activity dimension scores were slightly skewed and therefore presented as median and interquartile ranges (IQRs); moreover, the median, IQR, and minimum and maximum ranges of SGRQ dimensions and total scores are also illustrated.
Using Cronbach a for measuring internal consistency for the total score and each dimension (symptom, activity, and impact), we found a values of 0.94, 0.77, 0.91, 0.86, respectively. Total SGRQ had a highly statistically significant correlation with SGRQ symptoms, activity, and impact domains (0.64, 0.82, and 0.96, respectively). Symptoms domain had mild correlation with activity domain (0.41) and a moderate correlation with impact domain (0.61).
However, a better correlation was seen between activity and impact domains (0.7) (P < . 0001 for all these correlations). For test-retest reliability, the intraclass correlation coefficients for the total SGRQ and its dimensions (symptom, activity, and impact) were 0.97, 0.92, 0.94 and 0.95, respectively. There were small and statistically insignificant mean differences between scores at both completions, and the 95% limits of agreement of repeatability as shown in Bland-Altman plot.
The SGRQ total and dimensional scores showed high convergent validity correlating with different objective and subjective measures in respiratory research. The MRC dyspnea score, SF-36 PCS score, and SF-36 PCM score had the highest correlation with total and dimensional SGRQ scores. The maximum correlation was with both SF-36 scores and greater than the MRC dyspnea score (R2 = 0.5), suggesting that other important elements of ill health in these patients are being captured by the SGRQ.
Worse health status was associated with increasing severity of CPA; patients with severe CPA had worse health status than patients with mild or moderate disease. Moreover, the SGRQ showed significant discriminating ability in differentiating between all grades of shortness of breath. In addition, patients with more disease severity as justified by the respiratory physician scored consistently higher total and dimensional SGRQ scores.