A Comparison Of Two Versions Of The General Health Questionnaire Applied In A COPD Population

Bratås O, Grønning K and Forbord T

Objectives: To compare responsiveness, screening performances and internal consistency between version 20 and 12 of the General Health Questionnaire (GHQ), and concurrent validity of the GHQ-12 through agreement with The Hospital Anxiety and Depression Scale (HADS) in patients with Chronic Obstructive Pulmonary Disease (COPD) prior to and after pulmonary rehabilitation. Methods: Consecutive COPD patients with a mean age of 65, comprising 161 patients at baseline and 136 at 4-weeks follow-up. Responsiveness was analyzed through Paired Sample T-test, screening performances by Mc Nemar Test, internal consistency through Cronbach's alpha, and concurrent validity by The Bland-Altman technique. Results: The GHQ mean score was significantly reduced by the GHQ-20 and GHQ-12 after rehabilitation, with score differences of -4 and -3.9 using Likert format (p<0.001), and -3.3 and -1.9 using bimodal GHQ scoring (p<0.001), respectively. The two GHQ versions differed significantly in responsiveness through bimodal GHQ scoring (p<0.001). The prevalence of psychological distress was significantly reduced both by GHQ-20 and GHQ-12, using Likert format (48 to 30% and 64.3 to 40.3%, p<0.001, respectively), as well as bimodal GHQ scoring (36.4 to 19.3% and 41.1 to 21.7%, p<0.001, respectively). The two GHQ versions differed significantly in screening performances through Likert format, both at baseline (p<0.001) and after 4 weeks (p=0.004). Irrespective of scoring methods and time, the internal consistency was >0.9 for both versions. The concurrent validity of the GHQ-12 is considered low assessed through agreement with the HADS, with mean values of the difference of 4.1 at baseline (p<0.001) and 1.3 at follow-up (p=0.010). Conclusions: The two GHQ versions vary along with different scoring methods regarding responsiveness and screening performances. As for internal consistency, the observed differences were only marginal irrespective of scoring methods. Assessed through agreement with the HADS, the concurrent validity of the GHQ-12 is considered low.